Control of RNA Chain Elongation and Termination
Chapter
55
JOHN P. RICHARDSON and JACK GREENBLATT
Transcription of a DNA segment by action of RNA polymerase is an ordered, multistep process with three distinct phases —initiation, elongation, and termination. The mechanism of initiation is considered in chapter 54. This chapter is on elongation and termination.
The elongation of an RNA molecule formally begins with the condensation of the first 2 nucleotides (nt). However, this step and the subsequent 10 to 12 nucleotide addition steps are considered part of the promoter clearance stage of initiation rather than elongation because RNA polymerase readily releases nascent transcripts that are shorter than 13 residues and initiates new transcripts without dissociating from the promoter region. The transition to the elongation phase, which occurs when the nascent chain has 11 to 13 residues, is marked by three important changes in the transcription complex: a substantial increase in the stability of attachment of the nascent RNA (30, 154, 258), the dissociation of sigma factor from the RNA polymerase core (111), and movement of RNA polymerase away from the promoter region on the DNA (177).
Once RNA polymerase has entered the elongation phase, nucleotides are added to the 3' end of the nascent chain with an average rate of about 43 nt/s until a terminator is reached (183). During this phase the ternary complex of RNA, RNA polymerase, and DNA is highly stable to dissociation, with a half-life greater than several hours at most positions along the template. This stabilization allows the enzyme to transcribe a stretch of DNA as long as 20,000 bp, the length of DNA for some operons in Escherichia coli, without terminating prematurely. However, because of this stability, the enzyme needs special signals and often the assistance from special factors for termination of RNA synthesis when the end of a transcription unit is reached. A major aim of this chapter is to discuss the processes of elongation and termination, including the characteristics of the ternary complex during elongation, the mechanisms that lead to termination, and the ways in which the processes of elongation and termination are used to control the expression of genes.
The synthesis of RNA by action of a DNA-dependent RNA polymerase is called transcription because a DNA sequence is copied into a corresponding RNA sequence. The natural templates for RNA polymerases are double-stranded DNA. Through a particular gene region, only one strand is transcribed to yield a single-stranded RNA copy without permanently disrupting the double-strand structure of the DNA (83). Since an RNA molecule can form a stable double helix with a cDNA through H-bond pairing of RNA bases with DNA bases (178), it has long been presumed that the transcription process involves use of a temporarily separated single strand of DNA as a template for the assembly of an RNA by a mechanism analogous to that of DNA polymerases (33). The nucleotide addition reaction for both DNA and RNA polymerases involves a transfer of a nucleoside monophosphate from a nucleoside triphosphate (NTP) substrate to the 3' end of the growing polynucleotide, with PPi as the leaving group. The choice of the particular nucleotide to be added is presumably made by H-bond interactions between the base of the incoming NTP and the base of the DNA nucleotide on the 5' side of the nucleotide paired with the 3' end of the nascent polynucleotide. The recent finding that the structure of T7 RNA polymerase in and around its nucleotide addition site is very similar to the structure of the corresponding region of Klenow fragment DNA polymerase suggests that both enzymes could use similar mechanisms for nucleotide addition (249). However, for RNA polymerase to use a double-stranded DNA as a template, it needs to unwind the DNA for a short segment. It also needs to be able to displace the RNA chain from the template and to re-pair the two DNA strands. This locally unwound part of DNA, which is proposed in this basic model, is called the transcription bubble (276).
Several lines of evidence have indicated that the DNA is partially unwound in the elongation complex (276). However, a very good quantitative measure of the extent of unwinding came from a set of experiments by Gamper and Hearst (82), who measured the effect of the presence of E. coli RNA polymerase molecules in elongation complexes on the linking number of closed circular simian virus 40 DNA. The results indicated that the DNA is unwound by an amount equivalent to 18 ± 1 bp per elongating RNA polymerase molecule. This work thus validated the basic prediction of the transcription bubble model.
Elongation is a highly dynamic process involving the binding and addition of nucleotides and the translocation of RNA polymerase along the DNA template. To determine more about the structure of elongation complexes, particularly how the DNA interacts with the enzyme and the nascent RNA, procedures were developed to probe the structures of DNA and RNA in isolated elongation complexes in which the motion of the enzyme had been frozen at particular points on the DNA. The results of these experiments have provided further confirmation of the general feature of the transcription bubble. However, they have also revealed that the enzyme undergoes some structural transitions that were not predicted by the basic properties of the transcription bubble.
Since RNA polymerase pauses naturally at certain points on DNA for periods of up to 1 min, some studies were performed with complexes containing RNA polymerase at these natural pause sites. These complexes can be isolated by using a protocol that synchronizes initiation and allows synthesis by incubation with substrates for enough time for most RNA polymerase molecules to reach a pause site (99, 145). At that point the complexes are rapidly separated from the NTPs to prevent further elongation. Because the ternary complex may be in a special conformation at a natural pause site, other procedures were developed to isolate complexes in which RNA polymerase has been frozen at other points on the DNA. Usually, these alternative, "stalled" complexes are prepared by transcription of a DNA fragment in the absence of a single NTP that is not needed for transcription prior to residue 13 or higher—i.e., at a point after the enzyme has entered the elongation phase (154). Complexes isolated in these ways have been analyzed by a number of procedures, including DNase I and hydroxyl radical footprinting, chemical reactivity of unpaired residues in the DNA with diethyl pyrocarbonate and KMnO4, nuclease sensitivity of the RNA, and overall visual structure.
The way that RNA polymerase interacts with the DNA changes extensively when it moves away from a promoter. At a promoter, RNA polymerase makes contacts with the DNA that extend over 75 bp, from 20 bp ahead of the transcription start point to 56 bp behind (245), whereas in elongation complexes it contacts only 30 to 40 bp of the DNA (138). Another change is evident in the structures viewed by scanning force microscopy. In open complexes at the promoter, the RNA polymerase appears as a globule on a partially bent DNA molecule. However, in the images of an elongation complex, the DNA appears to be more acutely bent. Figure 1 shows a typical image of an elongation complex with RNA polymerase stalled 15 bp downstream from a start point. The mean bend angle for 35 such complexes was 92°, as compared with 54° for a similar number of open promoter complexes. Thus, coincident with a decrease in the overall contact between the enzyme and DNA is the formation of a sharper bend in the DNA.
A low-resolution model for the structure of E. coli RNA polymerase was proposed by Darst et al. (46) on the basis of results from electron crystallography. In this model, RNA polymerase has an irregular, globular shape, 10 by 10 by 14 nm in size, with a cleft at one end that resembles the hand-shaped clefts in the X-ray crystallographic structures of Klenow fragment DNA polymerase and T7 RNA polymerase. For a protein with these dimensions to make contacts that extend along 75 bp (25 nm) of DNA in the open promoter complex, some wrapping or folding of the DNA must occur, and these interactions could be responsible for the bending in DNA that is evident in most of the images of the promoter complexes. However, by this logic one would expect a lower bend angle with elongation complexes in which the protein makes contacts that extend only along 30 bp (or 10 nm) rather than the sharper bend angle that is observed. To understand the reason for these unexpected changes in the structure, further information will be needed from higher-resolution images of various transcription complexes.
When RNA polymerase translocates along the DNA during elongation, it does not have a single, fixed structural conformation, as was once thought. Instead, it has a conformation that is relatively fixed for most positions but undergoes a number of special conformational transitions when it passes through certain DNA sequences. This was first noted by Krummel and Chamberlin (139), who measured DNase I footprints in complexes paused at a number of successive points along a template. They found that the downstream boundary, the front edge of RNA polymerase, did not move forward as successive nucleotides were added to the RNA chain through certain regions of the template until a point was reached where it moved forward by about 10 bp. To explain these results, Chamberlin (32) proposed a specific model for the translocation of RNA polymerase along a DNA template. In this model the nucleotide addition site and part of the RNA product site are held in a fixed relationship with the part of the RNA polymerase that binds to the DNA behind the nucleotide addition site. During elongation these parts move along the template 1 bp at a time for each nucleotide added to the 3' end of the nascent RNA chain. However, the front edge of the enzyme remains tightly clamped at a specific site on the DNA until the distance between the front edge and the nucleotide addition site come to within 8 bp of each other. At that point the front edge releases its tight hold on the DNA, allowing it to slide forward by ∼12 bp, where it becomes reclamped, while the back edge and the nucleotide addition site maintain a fixed contact at their positions. Since the independent motion of the front and rear edges of the enzyme resemble the motion of an inchworm, this kind of RNA polymerase translocation is called inchworming.
Inchworm motion is a characteristic of transcription complexes passing through certain DNA regions but is not the motion used along most parts of a template. This was revealed by analyzing the positions of the front and trailing edges of RNA polymerase molecules in stalled complexes isolated at a large number of positions along a template. At most positions starting from about 20 bp downstream from the start point of transcription, the forward motion of the front edge and back edge of the enzyme is by single steps per nucleotide added (190, 266). Thus, the enzyme does use a monotonic form of progression through most DNA sequences. However, inchworm motion was found at certain specific locations. Since two of the specific locations were at the approach to a pause point (266) and at the approach to an intrinsic terminator (191), inchworm motion could have special roles in these important modulations to the normal elongation process. Also, since the original evidence for inchworm motion was obtained with complexes formed near the start points of transcription with two different templates, inchworming could also be a characteristic of promoter clearance in the earliest stage of elongation.
As expected in the transcription bubble model, RNA polymerase maintains a close and nearly continuous contact with the deoxyribose groups on the DNA backbone from about 13 nt behind to about 6 nt ahead of the nucleotide addition site on the template strand and 12 nt behind and 9 nt ahead of the corresponding residue on the other strand. This was evident from a hydroxyl radical footprint performed on a ternary complex stalled at 20 bp downstream from the T7A1 promoter (176). These contacts with the backbone presumably contribute to the very low rate of dissociation of RNA polymerase from DNA in the elongation complex but also allow the enzyme to move along the template in either single or multiple base pair steps.
The contacts of the enzyme with the backbone of the DNA strands could also help stabilize their separation so that one strand can serve as a template for assembly of the RNA chain, as expected in the transcription bubble model. Evidence showing that the two strands of DNA are not base paired with each other over a length of about 17 bp in transcription complexes has been obtained from reactions of bases with KMnO4, which can oxidize pyrimidines, especially thymines, via an attack from the face of the molecule (229), and with diethyl pyrocarbonate, which reacts with N-7 of purines, preferably on unpaired adenines (151). This approach has also provided evidence that is consistent with the proposal that an 8- to 10-nt segment of the template strand is paired with RNA.
Using several complexes isolated at some natural pause sites early in the λ late gene operon, Kainz and Roberts (130) found that T residues over a 17-bp region of the nontemplate strand reacted readily with KMnO4, starting at points very close to the residue corresponding to the base pair of the DNA at the pause point. T residues on the template strand also reacted if they were right at the base pair of the pause point and again in a segment of 4 nt starting at 15 nt behind the pause point. In other words, for a stretch of about 11 nt behind the pause points, the bases of the template strands are not very reactive with their reagents.
Lee and Landick (151) performed similar experiments with 12 different transcription complexes involving several DNA templates. They examined stalled complexes as well as those frozen at natural pause sites. Their set had a greater variety of reactivity than did the pause complexes of Kainz and Roberts. For instance, Lee and Landick found that from 14 to 22 residues were reactive on the nontemplate strand in different complexes, rather than just 17. They also noted that reactivity was generally greater on the nontemplate strand than on the template strand, except in the region of the nucleotide addition site, where residues on the template strand were more reactive than those at the corresponding position on the other strand. The region of high reactivity on the template strand was then followed consistently with about 8 nt of very low reactivity and then another six to nine residues of intermediate reactivity.
In sum, these results are consistent with the existence of a transcription bubble with about 17 bp of separated DNA that starts just preceding the nucleotide addition site and has from 8 to 12 nt on the template paired to RNA in a hybrid helix. However, because functional groups on the enzyme could also be masking bases on the template strand in the same region, these results do not prove that the final 8 to 12 residues of RNA are base paired to the DNA in the elongation complex. This point will be considered in further detail in later sections of this chapter.
E. coli RNA polymerase is a multisubunit enzyme. The core enzyme in the transcription complex has one each of the β and β' subunits and two of the α subunit. Genetic and biochemical analyses, along with sequence comparisons with other polymerases, have indicated that β has many of the components at the active site (132, 143). However, both β and β' contact the backbone of DNA in and around the start point of transcription in open promoter complexes (40, 245). Although it is not known whether these same subunits maintain similar contacts with the DNA in the elongation complex, there is good evidence that both subunits contribute to the active site and that various parts of nascent RNA molecules contact both subunits (290). Although the α subunit has a binding site for DNA that can make contact with an upstream element in some promoters (225), there is currently no evidence for direct contacts between DNA and the α subunits during the elongation process.
An expected property of RNA polymerases is the presence of a site or a cleft that acts to channel the RNA product away from the template strand. The extent of RNA that is wholly within the domain of the transcription complex was determined to be about 16 to 24 nt from the size of the fragments protected from mild RNase A digestion of complexes that had been stalled at random on DNA (140). This would presumably include RNA in the exit channel as well as the RNA that is paired with the template DNA strand. The degree of this partitioning is not known. Estimates that were based on reactivities of a photoactivatable group at the 5' end of a nascent transcript with DNA during the early steps of elongation suggested a length of 11 to 12 nt (110). However, this estimate was based on interactions that occur before the complex has made the transition to the committed state of elongation. When it was found that with high levels of RNase A all residues of the nascent transcript of isolated complexes were sensitive to within 3 nt of the 3' end of the chain (208), the possibility that the hybrid helix was as short as 3 bp long was considered likely and has prompted further probes. Lee and Landick (151) assayed the extent of RNase A sensitivity of the 3' end of the nascent transcripts in a number of isolated complexes and found that the RNA was generally more resistant in the final eight residues, with about the same sensitivity as a model RNA-DNA complex with an 8-bp hybrid helix. They also found that RNase VI, which is specific for RNA in helical segments, cleaves within the first 8 nt at the 3' end of RNA in some (but not all) isolated complexes. Thus, probes of RNA structure, as well as probes of the accessibility of bases in the template strand, appear to be consistent with a hybrid helix that extends for 8 bp.
As with the randomly halted elongation complexes analyzed by Kumar and Krakow (140), the RNA in elongation complexes halted at specific positions also shows a partial protection of the final 16 to 18 residues at the 3' end. Thus, if the final 8 residues are paired to DNA, this would leave about 8 to 10 residues that are protected by binding to the enzymes in what could be considered the product exit channel.
Photoactivatable reactive groups positioned at various parts of nascent RNA molecules in isolated ternary complexes have been used to determine which subunits or parts of subunits are in close proximity to the 3' ends of nascent chains and also which subunit could have the exit channel. Two separate studies using different photoprobes at the 3' ends of stalled complexes yielded very similar results. When a 5[(4-azidophenyl)thio]uridine residue was at the 3' end of a complex stalled at position 21 of the T7A1 transcription unit, it led to cross-linking of the RNA with both β and β' but primarily to β' (54, 290). When the RNA contained an 8-azidoadenosine residue at the 3' end of a 22-nt complex, the RNA became attached to the β' subunit, at a point between Met-932 and Trp-1020 (18). These two results suggest that both subunits contribute to the part of the RNA product site that is near the 3' end of the nascent transcripts. This is consistent with the evidence that both subunits contact regions of the DNA near the start point for transcription in open promoter complexes.
Complexes that contained the azidophenyl thiouridine analog at positions throughout the length of a paused transcript that was 80 residues long reacted with both the β and β' subunits (54). Since this nascent RNA would have reactive positions that could be contacting RNA polymerase outside the RNA exit site, the results do not distinguish whether this site is exclusively on one of the two subunits.
In E. coli, mRNA molecules from several operons are elongated at a rate of between 40 and 50 nt/s (21, 185, 264). Similar rates of growth were measured for synthesis of T7 RNA in vitro with E. coli RNA polymerase in reaction mixtures that contained intracellular levels of the four NTPs (51). These rates are about 20-fold slower than the rate with which a DNA chain is elongated during replication in E. coli. However, they are very close to the rate of mRNA translation of 16 codons per s (48 nt/s) (21), suggesting that the enzyme was designed to transcribe DNA at a rate that would allow close coupling of transcription with translation. However, E. coli RNA polymerase can synthesize RNA more rapidly in genes that are not translated. In rapidly growing cells, rRNA is synthesized at a rate of nearly 90 nt/s (264). This may reflect the maximum elongation rate achievable by E. coli RNA polymerase because if an enzyme had evolved that could polymerize RNA as fast as DNA polymerase III can synthesize DNA (i.e. 900 nt/s), the cell would not have needed multiple copies of the genes coding for rRNA (rDNA) to be able to synthesize ribosomes fast enough to sustain doubling of the cell mass every 20 min.
These rates that have been measured for transcription elongation in vivo and in vitro represent the averages through long transcription units. From studies of the progress of elongation of synchronously initiated RNA chains in vitro, it is known that the rate of addition of nucleotides can vary considerably from one position to another (45, 133, 144, 148, 153, 183, 224); RNA polymerase molecules distinctly pause during transcription of some sequences, and the pause points dominate the time needed to make a transcript. Under conditions in which it takes 60 s to synthesize a 400-nt transcript, 20 to 30 s is spent adding residues slowly at pause sites. This would mean that the rate of addition of nucleotides at sites where there is no obvious pausing would have to be at least double the average rate. It is possible in this respect that the reason why rRNA molecules are synthesized at nearly twice the average rate as mRNA molecules is that the rDNA sequences are transcribed with little, if any, pausing.
A functional role for pausing in the synthesis of an mRNA is to coordinate transcription with translation. This role is suggested by the finding that strong pause sites downstream from the ribosome-binding sites are a conserved feature in the attenuator regions of amino acid biosynthetic operons (146).
From analyses of the sequence components that contribute to pausing, five distinct features have been identified (36, 153). A key feature of many pause sites on a DNA is the ability of the transcript to form a short stem-loop structure (hairpin) that has its 3' end about 10 to 11 nt from the 3' end of the RNA in a paused complex. A role for the stem structure has been inferred from the finding that a mutation that would disrupt the pairing of two bases in the stem reduces the extent of pausing, but this lost function can be recovered by a compensatory mutation that restores pairing. The segment of DNA ahead of the RNA polymerase also affects pausing. However, the feature of this downstream sequence that correlates with efficiency of pausing has not yet been recognized. Similarly, changes in residues in the region of DNA from the pause points back about 11 nt influence the extent of pausing. This includes the region of the RNA that could form the 8-bp hybrid helix with the template. Again, it is not known how these nucleotides affect the pausing nor whether the 3' end of the RNA is paired with the DNA in the pause conformation. A fourth feature concerns the nucleotide to be added at the pause points: many natural pauses occur at a point when a G residue is to be added to a pyrimidine residue; mutations that change the 3' end of the nascent RNA at the pause point affect the position and duration of the pausing. Structural differences of the nucleotides in the donor and acceptor sites for the nucleotide addition reaction could readily influence the rate of the addition reaction. A fifth feature has been identified in a natural pause point that allows Q protein to form an antitermination complex with RNA polymerase starting at the λ late-gene promoter. Using heteroduplex templates, Ring and Roberts (217) found that the bases on the nontemplate strand that would be unpaired in the elongation complex at the pause, but in the region close to where the DNA becomes paired again at the following edge of the transcription bubble, could influence the duration of the pause. This is an interesting finding, as the studies with reactivity of the bases in the region have suggested that the enzyme does not normally make base-specific contacts with any part of the separated nontemplate strand. Perhaps a certain sequence has a special structure that is recognized in the single-stranded form or the enzyme is able to bind tightly enough to certain bases to retard the progression of the enzyme along the DNA.
Wang et al. (266) have used exonuclease III to probe the position of RNA polymerase on a DNA template and RNase V1 to probe the extent of protection of nascent RNA in transcription complexes stalled at several positions on DNA leading into the transcriptional pause site of the his attenuator. They found that the distance between the active site and the front edge of the polymerase, as judged by the exonuclease III block point, remains roughly constant for most positions along the template until the active site is 8 bp before the pause point. Their results thus confirm the observation of Nudler et al. (190) that translocation is stepwise and continuous along most points of the template. However, starting when the active site was positioned 8 bp prior to the pause point, the front edge of the polymerase ceased to move forward with each nucleotide added until the active site reached the pause site. At that point, the front edge moved forward by 8 bp. Thus, a discontinuous, inchworm motion corresponded precisely with the approach to a pause site on the his attenuator DNA. Concomitant with this discontinuity in the motion along the DNA was a discontinuity in the accessibility of the nascent RNA to RNase V1. In complexes arrested up to about 6 bp before the pause point, RNase V1 could cleave to within 18 nt of the 3' end of the nascent RNA. Starting at that point, the addition of each new nucleotide leads to an increase in the extent protected up to a total of 22 nt protected in complexes arrested at 2 bp before the pause point. With the final two nucleotide addition steps, there was an abrupt change to a structure in which about 35 nt were protected.
These results have been interpreted in terms of the formation of a strained transcription complex that relaxes as the enzyme reaches the pause point and is associated with a major structural transition in the RNA. On the basis of sequence features that correlate with pausing sites and of structural studies of the RNA and DNA in isolated transcription complexes at pause points and at points leading up to a pause point, Chan and Landick have suggested a general model that can account for pausing at sites that involve formation of a stem-loop structure in the RNA (36). The features of the model are shown in Fig. 2A. In the model the final 8 to 10 residues at the 3' end of the RNA either are involved in base-paired interactions with the separated template strand of the DNA or are held closely apposed to the template in a proximal RNA-binding site with at least the last 2 residues paired with their template bases. Under the normal processive elongation mode, another 8 to 10 residues would be held in an extended, single-stranded conformation in the exit channel, the distal RNA product site, by H bonds and ionic bonds with the sugar-phosphate backbone of these residues. However, as the enzyme approaches a pause site, a structure or sequence on the DNA near the pause point—either just downstream or just upstream—somehow holds back the forward motion of the front edge of the polymerase, but without blocking nucleotide addition. The residues added at this stage are accommodated by a motion of the active site along an unoccupied portion of the proximal RNA-binding site. The addition of these nucleotides, while forward motion is blocked, induces a strain that is relieved by an isomerization in the complex that includes the formation of the RNA stem, which makes new contacts in the RNA exit site. The new contact with the stem could put the polymerase into a conformation that no longer can add nucleotides to the 3' end of the nascent RNA, possibly because the 3' end is pulled away from the nucleotide addition site. The resumption of elongation would then require another isomerization that would release the RNA from this inactive configuration so that nucleotide addition could resume.
Chan and Landick (35) found that the extent of pausing at the hisL pause site is very sensitive to the ionic conditions —much more sensitive than the general rate of elongation. Hence, they have proposed that electrostatic interactions of the stem-loop with a site in the enzyme is responsible for holding the complex in the paused configuration. The role of the other sequence features that are associated with pausing could be to favor conditions that allow the major conformational change that puts the complex into the paused mode. Since not all pauses are at points where residues in the RNA will form obvious stable stems within the exit site, some other features in the transcription complex could allow a conformational change that is not stabilized by the stem structure. Perhaps a combination of the other sequence features contribute to retard elongation to create a significant pause.
RNA polymerase molecules do not behave homogeneously with respect to pause sites (34, 133). Some molecules appear to pass by a pause site with little or no delay, while others stop moving. This suggests that distinct conformational changes occur with the pause. For those that do pause, the kinetics for resumption fit well to a first-order reaction. This suggests that the resumption of elongation occurs via an isomerization reaction. An individual pause site is thus characterized by two parameters: the efficiency of pausing and the half-life for resumption of elongation. Each site can differ in both parameters.
One functional characteristic of a true pause site is that eventually all RNA polymerase molecules that are paused there will resume elongation. This implies that the isomerization that put the enzyme into the paused configuration is reversible. However, there are points on DNA where changes occur in the RNA polymerase molecules that prevent them from resuming elongation (11, 138). These sites are operationally distinct from pause sites because the enzyme forms a dead-end complex. They also differ from termination sites by not allowing release of the RNA. However, they are similar to both pause and termination sites in that they do not function with 100% efficiency. Usually only a certain fraction of polymerase molecules become arrested upon transcription into such a site, which suggests that the arrest is the result of a conformational change that is triggered by certain signals. Since the presence of an RNA polymerase permanently arrested on the DNA is a type of functional DNA damage, the cell has devised mechanisms that either prevent the arrest from occurring or reactivate the enzyme in the arrested complexes or remove the damaged enzyme. These mechanisms are considered in a later section.
A large number of mutant forms of RNA polymerase that have altered kinetics of elongation have been analyzed (69, 103, 128, 186, 281). Many of these mutants were recognized because they also alter the efficiency of termination. These mutants are often also resistant to the drug rifampin, a specific inhibitor of bacterial RNA polymerases. Rifampin binds to the β subunit of RNA polymerase either as the free protein or in its complex with DNA, but it has no effect on enzyme molecules that are in the process of synthesizing an RNA chain (246). In the presence of rifampin, RNA polymerase is able to initiate RNA chains but is unable to make products that are longer than three residues (129). This mode of inhibition suggests that rifampin binds to the part of the RNA product site that interacts with the nascent RNA 4 to 5 nt from the 3' end. Thus, a mutational change in the RNA polymerase that affects the binding of rifampin might reasonably affect how RNA polymerase interacts with the nascent RNA in that part of the chain and thus affect the processes of elongation and termination.
Several different rifampin-resistant mutants have been analyzed, and nearly all affect termination of transcription at several different terminators (128). All mutations that confer resistance to rifampin are in the rpoB gene, which encodes the β subunit, and most are localized in a limited region of that gene (126). One mutant enzyme, RpoB8, has a β subunit with a proline at residue 513 rather than a glutamine. RpoB8 RNA polymerase has a reduced elongation rate, has an increased Km for purine nucleotide triphosphate (127), and pauses significantly longer at the trpL pause site (69). Thus, the single change apparently has affected the nucleotide-binding site as well as the rifampin-binding site, despite the presumed separation of the two.
Landick and his coworkers (143) have isolated another RNA polymerase mutant (RpoB5101) that fails to pause at the trpL pause site. This mutant differs from RpoB8 in that it has a normal Km for NTPs during elongation and it has no defect in transcribing a simple DNA polymer. Its defect must be specifically affecting its ability to enter the pause configuration. It has two amino acid changes: serine instead of proline at position 560 and isoleucine instead of threonine at position 563. Although these changes are not far from that of RpoB8 and do not confer resistance to rifampin, they map in a region where rif resistant alleles are found. These results thus suggest that residues in a limited region that include residues 513, 560, and 563 form part of the RNA chain elongation site and are critical in controlling the elongation state of the enzyme.
Streptolydigin, another antibacterial drug, inhibits E. coli RNA polymerase by interfering with nucleotide addition (31); thus, it decreases the rate of elongation. Since it can bind to and inhibit RNA polymerase during elongation, its binding site does not overlap with that of the nascent RNA. Mutations in the rpoB gene that confer resistance to streptolydigin affect residues that are near those that are involved with rifampin sensitivity (114). However, they are in a small region of the rpoB gene that can be deleted without loss of function, and mutant enzymes that are resistant to streptolydigin are normal or nearly normal in their function. Thus, the drug-binding site for this inhibitor appears to be in a fold or loop that is not critical for polymerase function, although binding of the drug to this site does interfere with nucleotide addition.
E. coli has several protein factors that can modulate the rate of RNA chain synthesis either generally or with very specific transcription units. One of these is the product of the nusA gene, a protein with an M r of about 55,000 (123). Another is NusG, an abundant protein with an M r of 21,000 (28, 155).
NusA acts to prolong pausing at certain natural pause sites (64, 70, 133, 136, 148) and has a general inhibitory effect on chain elongation that can be overcome by high levels of NTPs (231). Pure NusA protein binds reversibly to core RNA polymerase, either as a free enzyme or as part of a ternary transcription complex, but does not bind to holoenzyme (93, 105, 118, 232). Thus, it replaces sigma factor during elongation and could occupy the same binding site as sigma on the core enzyme.
NusA has two distinct effects on elongation; it increases the Ks for the NTP substrates for RNA polymerase and it accentuates pausing at certain sites (231). The first of these effects is general and leads to a retardation of chain growth when the concentration of NTPs is low. The mechanism of the alteration is not known; one likely possibility is that the binding of NusA to core RNA polymerase allosterically modifies the structure of the nucleotide addition site. The second effect is specific for certain pause sites and is thus dependent upon recognition of a sequence-encoded signal. The pause sites that are specifically affected by NusA all belong to that class in which the RNA transcript has the potential to form a stable stem-loop as part of the pause complex. However, not all pause sites that have a stem-loop signal feature are affected equally by NusA. This observation has led to the hypothesis that this discrimination arises because of a direct interaction of NusA with the nascent RNA in the RNA exit domain. However, the sequence determinants for this hypothetical discrimination are not yet evident. In an alternative model, the discrimination is achieved by a stabilization of the binding of NusA to RNA polymerase that could occur when the enzyme enters certain pause conformations. Landick and Yanofsky (145) found that NusA enhanced the protection of residues from T1 RNase cleavage in the stem structure of nascent RNA in a complex isolated with RNA polymerase at the pause site at bp 93 in trpL DNA. However, this result could be explained by indirect as well as direct effects.
Evidence that NusA actually contacts specific regions of the nascent RNA has come from the finding that it becomes cross-linked to RNA containing photoactivatable groups in certain positions upon illumination of isolated pause site complexes (162). However, this reaction is very sensitive to the position of the reactive group (55), which suggests that NusA is contacting specific regions on RNA. The exact positions of the contact sites are not known yet. Also, the presence of NusA alters the reactivity of a photoactivated group at the 3' end of RNA in a stalled complex, increasing particularly the reactivity of the group with the β subunit (290). This change is likely related to the general effect that NusA has on the Ks for nucleotide addition.
NusA has an essential function in E. coli (43). Before its identification as the product of the nusA gene (106), the protein was isolated as a factor, called L, that enhanced the yield of synthesis of large proteins in a partially fractionated, coupled transcription-translation system (141).
One likely function of NusA is to retard continued elongation of RNA polymerase at critical control points in an operon in order to keep translation of the mRNA closely coupled to transcription. Many genes contain latent Rho-dependent terminators that are quite effective when the transcripts are not being translated but are functionally masked by normal translation (226). By causing RNA polymerase to pause at a point between the start point of translation and the first intragenic, Rho-dependent terminator, NusA would allow the ribosome to bind and initiate translation before the RNA polymerase passes into the termination region with an unprotected nascent RNA. The finding that NusA is no longer essential in strains of E. coli with mutations in the rho gene that greatly reduce the termination activity of Rho factor (291) is consistent with the notion that its major function is to ensure efficient coupling of transcription with translation.
The use of NusA to couple transcription with translation also occurs in the functioning of attenuators for amino acid biosynthetic operons. NusA enhances pausing at a site between the translation initiation codon and the transcriptional terminator in the trp and his operon leaders (34, 64, 144). Since the functioning of these attenuators depends on the coupling of transcription with translation of the leader mRNA section, NusA is likely serving a critical role as that coupling factor. This role is also consistent with the findings that translation of the nascent leader transcript in vitro relieves the transcriptional pause (142).
NusA also plays an essential role, along with some other host proteins, in mediating antitermination by the product of bacteriophage λ gene N (76). Its name, an acronym for N utilization substance, derives from this function, as a mutation in nusA was isolated that makes the host unable to use λ gene N to activate the expression of the λ delayed early genes. NusA, other host proteins, and λ N combine to form a cis-acting antitermination complex. Further details about this antitermination mechanism are considered in a later section.
NusG has an action opposite that of NusA; it enhances the rate of transcription elongation. The overall effect on elongation is not very large, as depletion of NusG in vivo increased the time for β-galactosidase induction by only ∼20% (28). However, from studies of its effects on the kinetics of chain growth in vitro, NusG was found to reduce pausing, and for one pause site, at least, it acted to reduce the half-life of the pause by a factor of 2 (28). NusG is known to interact with RNA polymerase (155). Thus, its presence in the elongation complex could facilitate the isomerization step that converts an enzyme molecule in a paused conformation back into its elongation conformation.
Two E. coli protein factors, called GreA and GreB, induce cleavages near the 3' end of RNA molecules in stalled or arrested ternary transcription complexes (19, 20). They also cause cleavages near the 3' end of isolated RNA molecules in the presence of E. coli RNA polymerase (6). These proteins are considered elongation factors because they suppress the process of transcriptional arrest with RNA polymerase. In the ternary complexes, the cleaved 3'-end fragments dissociate and RNA polymerase is able to resume transcriptional elongation from the new 3' terminus of the cleaved transcript.
GreA and GreB have the same number of amino acid residues and are similar in primary structure (20). A coiled-coil region of GreA consists of amphipathic α helices, with a basic patch that contacts the nascent RNA (251). Although GreA and GreB cause RNA to be cleaved, they bear no structural resemblance to known endonucleases or exonucleases. This characteristic, along with the sensitivity of the cleavage to inhibitors of RNA polymerase and some other properties, points to a role for these factors in inducing cleavage catalyzed by functional groups within the RNA poly-merase itself. Consistent with this idea, RNA polymerase that is purified from a greA greB mutant strain can cleave a nascent transcript at mildly alkaline pH in the absence of GreA or GreB (198).
The two factors also have some subtle but significant differences in the types of transcript cleavage that are induced. With GreA, the 3'-end fragments are usually di- or trinucleotides, while with GreB the fragments are longer oligonucleotides, up to a length of nine residues (20). For GreA to function, it has to interact with the transcriptional complex before elongation is arrested. In contrast, GreB can reactivate arrested complexes. It causes cleavage of 3'-end fragments in these complexes in a way that allows RNA polymerase to resume elongation of the 5' fragment. However, despite these differences, the outcomes are similar; they afford a substantial suppression of the process of transcriptional arrest.
The formation of dead-end, arrested complexes occurs under certain circumstances during transcription of DNA in vitro with RNA polymerase that is devoid of GreA and GreB (20, 63). One naturally occurring reaction of RNA polymerase that is known to cause formation of arrested complexes is misincorporation of a noncomplementary residue during transcription (63). Since the fidelity of the transcription process is not perfect, misincorporation of residues is likely to occur frequently enough—perhaps once per every 1,000 residues—to be a significant source of damage of DNA function from the formation of arrested complexes. Thus, the Gre proteins prevent the damage either by removing misincorporated 3'-end residues before formation of the arrested state or by reactivating complexes that have become arrested because of misincorporation.
DNA molecules can be damaged by reaction with certain molecules or from reactions that are induced by absorption of UV light, and many types of damage will arrest elongation during transcription (177). Because the presence of an arrested RNA polymerase at the site of damage would prevent access by the enzymes that recognize and initiate repair of damaged DNA, E. coli has a protein known as the transcription repair coupling factor (TRCF) that serves to dissociate these arrested complexes (238). This protein, the product of the mfd gene (239), consists of 1,148 amino acid residues with a segment that is similar to the UvrA-binding region of UvrB, another segment that has motifs that are found in known helicases, and a third segment containing four leucines at seven-amino-acid intervals.
In a defined transcriptional system, E. coli TRCF specifically acts to dissociate the ternary complex arrested at a DNA lesion (239). Hence, TRCF is not a transcriptional elongation factor but actually a transcriptional termination factor that is specific for arrested complexes. A possible model for its mechanism involves the use of the leucine repeats for interaction with RNA polymerase in the arrested complex, the use of the helicase motifs in dissociating the nascent RNA, and the use of the UvrA-binding motif to recruit a UuvA-UuvB complex to the site of damage after TRCF has mediated the release of RNA polymerase. The activity of UuvB then can initiate the series of reactions that lead to repair of the damage in the DNA (240).
One possible source of disruption of transcriptional elongation is an encounter with a DNA replication fork. Replication forks appear to be able to pass by RNA polymerase in stalled, transcriptional elongation complexes without necessarily dissociating the complex. This has been shown from some in vitro model studies of reactions with the T4 DNA replication apparatus on DNAs with E. coli RNA polymerase complexes oriented in either direction with respect to the motion of the replication fork (160, 161). Perhaps the abilities of RNA polymerase to make contacts with DNA on both sides of the transcription bubble and to bind RNA tightly in the RNA product site are mechanistic features that can accommodate the passage of a replication fork. By having two distinct, sterically separated sites for binding DNA, RNA polymerase could maintain a contact with a position on the DNA at a distal site when the proximal contact is released to allow the replication fork apparatus (DNA helicase, DNA polymerase, processity clamp, and other proteins) to enter the transcription bubble region. After the proximal contact on the replicated DNA segment is reestablished, contact at the distal site could be released to allow passage of the fork from the intervening DNA. Presumably, interaction of the 3' end of the nascent transcript with the DNA template strand is used to reestablish the correct position of the 3' end for continued transcriptional elongation. Thus, a critical function is the ability of RNA polymerase to hold the RNA tightly in the RNA product site (6, 32) during disruption of the transcription bubble by the replication process.
Termination of transcription occurs when a nascent RNA is released from its complex with RNA polymerase and DNA template. This can happen spontaneously at certain sequences in the DNA or be mediated by action of a factor (213). Just as the stability of the ternary transcription complex is important for the efficient copying of a gene sequence, the ability of the enzyme to cease further elongation and to release the transcript at the end of the gene is important for orderly expression. This is particularly true for transcription of bacterial genomic DNA in which the genes are organized with very short sequence segments between the sequences that encode functional products.
The interactions between RNA polymerase, DNA, and the nascent RNA have evolved so that almost all sequences can be transcribed with a very low probability for release of the nascent chain. However, there are sequences where the probability of release is high enough to be competitive with continued elongation (273). These sites function as intrinsic terminators and are specified by a limited set of sequences. These terminators represent a major class of functionally active termination sites. In addition, E. coli, Salmonella typhimurium (official designation, Salmonella enterica serovar Typhimurium), and probably all other bacteria use another mechanism that involves a different set of sequence signals. Since these other terminators involve the function of a protein factor called Rho, they are called Rho-dependent terminators (203). Currently all known terminators in E. coli can be classed as either intrinsic or Rho dependent.
A large number of intrinsic terminators have been identified, and analyses of their sequences have revealed some very distinct motifs (22, 29, 223). These are characterized by about 20 bp of a G+C-rich sequence with an interrupted dyad symmetry preceding (in the direction of transcription) a sequence of about 8 bp with a run of dA residues on the template strand. There is considerable evidence that these sequence motifs are important parts of the terminator. However, other features of the sequence within 30 bp of the transcription stop point that are not evident as an obvious consensus sequence are also important. The sequence following the run of T·A base pairs is in this category. Changes in the sequence 3 to 5 bp downstream from the stop point for the T7 early gene termination (T7Te) can cause the efficiency of the terminator to drop from 65 to 10% (262).
A terminator is operationally defined as a point where the rate of release of an RNA transcript, k release, is greater than the rate of addition of the next nucleotide, k forward (265). The sequences at an intrinsic site must have features that can allow a great increase in k release and/or a great decrease in k forward, when compared with "average" sequences. The crucial role of the determinants of the k forward rate in the termination process was shown by the evidence that the efficiency of utilization of a number of terminators is strongly dependent on the concentration of the NTP that is the substrate for the next nucleotide to be added at a specific stop point and on mutational alterations of the RNA polymerase that affect the intrinsic catalytic activity of nucleotide addition (174).
At least three aspects of the terminator sequences have features that are likely to contribute significantly to those changes. The first is the G+C-rich sequence with the interrupted dyad symmetry. That sequence would have been transcribed by an RNA polymerase that has reached the termination end point and gives rise to a segment of RNA that would be capable of forming a very stable stem-loop secondary structure. Whether the structure actually forms in the nascent RNA in a transcription complex at a termination end point has not been demonstrated, but there is abundant indirect evidence that it does. First and foremost, mutants that change the sequence in a way that affect the pairing of the stem affect termination (41, 168). For instance, a change that interrupts the pairing in the stem decreases the efficiency of termination, but most of the efficiency can be recovered if a second change restores the pairing. Second, the mutational change has its effect only if it is in the template strand, thus ruling out possible models which involve a structure in the nontemplate DNA strand rather than the transcript itself as the feature that is detected (227, 277). Third, there is a strong correlation between the predicted stability of part of the RNA structure and termination efficiency of variants with changes in the stem. However, this correlation is not absolute, as certain mutants of the tR2 terminator which encode RNAs that should have equivalent stem stabilities can have very different termination efficiencies. These results suggest that the actual sequence of the RNA that forms a stem also has a role in determining the efficiency of termination (41). Fourth, the presence of DNA oligonucleotides that are complementary to the 5' strand of the stem specifically reduces efficiency of termination at an intrinsic terminator (70), presumably because the binding of the oligonucleotide preempts formation of the stem.
Mechanistically, how might the formation of a stem structure in the nascent RNA contribute to the termination process? As mentioned earlier, some pause sites also occur at points where the nascent transcript could form a stem-loop structure. Because of the coincidence of the presence of such a structure at both kinds of sites and because pausing of elongation would provide time for other changes in structure that allow release of the transcript, it seemed likely that these stems associated with termination are functioning primarily to slow down the step time for elongation (k forward) at the termination site. Although this conjecture has not been directly demonstrated, it is consistent with the observation that the elongation factor NusA, which increases pausing by RNA polymerase, increases the efficiencies in vitro of many intrinsic terminators (64, 99, 108, 233). The distance from the end of the stem structure to the 3' end of the RNA is slightly different for a terminated RNA. With most terminators the base of the G+C-rich stem is about 7 to 9 nt from the 3' end of the released transcript, whereas with transcripts at pause sites the distance is 11 to 12 nt. However, because the stems probably form by a pairing process that is initiated across the loop, a stem with a pause-like configuration would form before being extended to the full-length stem of the terminated transcript, and this intermediate stem might be sufficient —along with other pause-inducing signals—to initiate the pause process.
The ability to form a stem structure could also contribute to the termination process by weakening the overall interaction between the nascent RNA and the exit region of RNA polymerase. Arndt and Chamberlin (11) have shown that ternary complexes with RNA polymerase arrested or paused at certain points that are not termination stop points dissociate at rates that are significantly higher than at other points. In all cases, these low-stability pause sites are at points where a short stem could form in the nascent RNA, with the stem ending within 6 to 10 residues of the 3' end of the transcript in the paused complex. This result suggests that formation of the stem leads to instability of the ternary complex. Since the positions of these partial "destabilizing" stems are at the positions of the stems in terminated transcripts, the termination RNA stems are likely to contribute to an increase in k release as well as to a decrease in k forward at the termination stop point.
The second aspect of the sequence for an intrinsic terminator is the run of dA residues in the template strand. Since termination stop points usually occur after several of the dA residues have been transcribed, the RNA molecules will have a run of U residues at their 3' ends. Again, the importance of the run of dA residues has been tested by mutational analysis (168). In the conventional view of the transcriptional elongation complexes, most or all of these eight terminal U residues would be base paired with the dA residues in the template strand. However, as pointed out in the section on elongation, many of the U residues may already be displaced from the template and within the RNA exit site of RNA polymerase instead (208).
There is excellent physicochemical evidence that a hybrid helix consisting of rU residues paired to dA residues is significantly less stable than most other hybrid helices (170). The possible reason is that a run of dA residues does not fit well into A-form double helices, the exclusive form that is taken up by helices that contain at least one RNA strand. Thus, poor pairing of the 3' end of an RNA to the template could readily contribute to an increase in k release. This basic property of RNA-DNA duplex helices is a compelling reason to believe that the normal elongation complex does involve a hybrid helix of 8 bp or more.
The third aspect of the intrinsic terminators is the sequence ahead of the polymerase at the termination stop point (207, 262). Although this sequence is not transcribed to become part of the terminated transcript, it is within or just ahead of the contact point between the enzyme and the DNA in the transcription complex at the termination site. Since downstream sequences are known to affect pausing at nontermination sites (152), a major role could be to contribute to a decrease in k forward. Those sequences could be influencing the unwinding of the DNA or just the progression of the enzyme along the DNA. Alternatively, since downstream sequences are within the contact site, there could be some variations in the stability of the binding of RNA polymerase related to sequence variations in the contact points. Thus, the downstream sequence could conceivably increase k release at a termination stop point.
The effects that a termination sequence and components of a termination sequence have on the mode of translocation of RNA polymerase along the DNA template have been examined by Nudler et al. (191). When exonuclease III was used to probe the positions of the front edge of RNA polymerase with respect to the nucleotide addition site in complexes arrested at various points, the typical distance of 18 bp that was characteristic of most steps in elongation was found to be less when the active site was within 9 bp of the release point, indicating a retardation of the forward motion of the front edge in the termination region. Since complexes with RNA polymerase at the stop point sites were not stable to isolation, it was not possible to probe the position of the front edge at those points without altering the complexes. By incorporating inosine residues at positions of the stem of the RNA that would form at the termination site, it was possible to isolate complexes that did not dissociate rapidly. With these, the front edge of the enzyme had moved forward to give the 18-bp distance that is characteristic of normal elongation positions. This result thus suggests that a discontinuous jump of the polymerase to a relaxed conformation is a mechanistic feature associated with termination.
A test of the relative roles of the two major parts of a terminator, the run of dA residues in the template strand and the upstream sequence that encodes the part of the RNA that forms the stem structure, showed that the run of dA residues was the one responsible for the temporary arrest of the forward motion of the front edge of the polymerase while the part encoding the stem structure in the RNA was crucial for the release of the RNA but did not cause a discontinuity in the elongation cycle. Thus, two properties of the run of dA residues on the template strand could be making important contributions to termination—the ability of this segment of the DNA to trigger a discontinuity in translocation and the poor pairing between the template dA residues and the U residues at the 3' end of the RNA transcript.
Although an intrinsic terminator and a pause site have two features that are very similar—the presence of a sequence that encodes an RNA stem and a sequence that causes a discontinuity in the translocation of the front edge of RNA polymerase—they differ in two key features. One is the spacing between the end of the RNA stem structure and the point where the discontinuous leap occurs. This is 8 bp for a terminator and 11 bp for a pause site. Thus, the spacing of the sequence that causes the discontinuity with respect to the sequences that encode the RNA stem structure appears to be crucial. The other difference is that the terminator encodes an RNA that ends with a run of several U residues. Because of the shorter spacing in a terminator, the stem in the RNA may form too soon to allow interactions with the site that stabilizes the ternary complex in the paused conformation. Instead, by forming earlier, the stem pulls the RNA away from the contacts that normally stabilize the attachment in the RNA exit site. A diagram for this is shown in Fig. 2B. This change and a weaker base pair interaction between the nascent RNA and the DNA that is characteristic of the rU-dA pairs could be the features that allow rapid dissociation of RNA at the terminator.
The overall process of transcription termination also includes the release of the RNA polymerase. Attempts to determine whether that step is simultaneous with release of RNA or occurs subsequently have given ambiguous results (10). Using a kinetic approach, however, Arndt and Chamberlin (12) were able to show that the addition of excess sigma factor significantly increases the rate of enzyme recycling in a system that involves synthesis of a 160-nt transcript from a DNA template with a strong promoter and efficient intrinsic terminator. This result fits well with a model in which sigma is directly involved in catalyzing release of core RNA polymerase from DNA at a step after release of the RNA. Thus, in its general capacity of decreasing the affinity of core for nonpromoter DNA sequences (115), sigma factor can also be considered a termination factor.
Sequence Elements.
Rho-dependent terminators consist of two distinct, but partially overlapping, parts that together extend over 150 to 200 bp of DNA (147, 149, 211, 283). One part is the region where termination occurs, the transcription stop point (tsp) region, while the other is an essential upstream region called rut. Both regions are fairly broad. The stop points of a Rho-dependent terminator are often spread over 100 bp of DNA in clusters of preferred points called subsites (147, 182, 274). This is in contrast to the stop points for an intrinsic terminator, which are concisely localized over a region of only 2 to 3 bp. The rut region adjoins the tsp region and has a minimum size of 85 bp (113, 184). The sequences that specify a Rho-dependent terminator are very diverse and do not conform to a simple consensus. However, there are some distinct limitations in the types of sequences of a rut region, and the positions of the preferred stop points within the tsp region are governed by sequences within the region.
Several lines of evidence indicate that rut sequences are recognized on the RNA transcript and represent the binding site for Rho. First, Rho factor is an RNA-binding protein that has a distinct preference for the type of RNA molecule to which it will bind (81, 209). Second, its preference for binding to the transcripts correlates with the presence of functional rut sites in the transcript; it binds well to transcripts terminated by Rho action but not to isolated nascent transcripts that have not been extended to the end of the rut region nor to full-length transcripts of DNA in which the rut site has been deleted (65). Third, DNA oligonucleotides that are complementary to the rut segment of the transcript specifically block Rho-mediated termination during transcription, presumably because these oligonucleotides can compete with Rho for binding to the rut sequence on the transcript and block the binding of Rho to the isolated transcript (39).
Studies on its general preference for RNA show that Rho has a very strong affinity for poly(C), a moderately high affinity for single-stranded RNAs without C residues, and a much lower affinity for RNA molecules in which most of the bases are paired with other bases (81). Analyses of rut regions and likely rut regions have shown that indeed they generally consist of stretches of segments that are not likely to pair with other segments of the RNA and that have a higher than average proportion of C residues (181). Also, a sequence characteristic that correlates well with a large number of Rho-dependent terminators is a compositional bias of a relatively low G and high C content in the part of the transcript just preceding the 3' ends of terminated transcripts (5). The low level of G is important because of the propensity of G residues for pairing with other residues and thus contributing to double-helical secondary structures in the RNA. Overall, the general view that has emerged of a rut site is that it encodes RNA segments of at least 85 nt in length that are largely single stranded and contain some C residues.
The fact that Rho has such a high affinity for RNA with unpaired cytidine residues has prompted some systematic studies of the role of cytidine residues at functional rut sites. In one study a natural rut site was altered to change several of the C residues to U residues. This work indicated that no specific cytidine residue appeared to be required and at least 13 of the 28 cytidine residues in the 104-nt trp t' rut region can be mutated with very little loss of termination efficiency (284). However, a derivative with only 11 changes was significantly less efficient than others with a similar number of alternative changes, implying a certain hierarchy in the pattern and numbers of residues. An alternative approach involved starting with a template that encoded an inactive AU-rich region upstream from the tsp region of tR1 of bacteriophage λ and determining the number and distribution of C residues that were present in mutational derivatives that had functional terminators (112). The results showed that the presence of from five to six cytidine residues, preferably spread through an 80-nt segment, was able to activate efficient transcription termination and that no obvious clustering or regular spacing was necessary.
The sequence requirements for the tsp region are even less well defined. Recent experiments have indicated that several different nontermination sequences placed downstream from a good rut sequence each became a tsp region with a characteristic set of new transcription stop points (L. V. Richardson and J. P. Richardson, unpublished data). Studies of the kinetics of transcription elongation through the tsp region in the absence of Rho factor revealed that the distribution of pause points correlates very closely with the distribution of stop points from Rho-dependent termination (148, 183). This correlation suggests that Rho causes RNA polymerase to terminate preferentially at natural pause sites downstream from a rut region. Furthermore, this action of Rho appears to be kinetically matched to the motion of RNA polymerase along the template (125). This was inferred from the finding that the poor termination activity of a slow-acting (as indicated by low ATP hydrolysis activity) Rho could be overcome by using a mutant RNA polymerase with a slower elongation rate or by slowing the rate of motion of wild-type RNA polymerase with low levels of substrates. These results suggest that Rho is acting primarily as an RNA release factor, with its specificity being determined by where it is able to bind to the RNA and where the RNA polymerase is on the DNA after the bound Rho has started to act. These positions on the template are determined by the rules that govern transcriptional elongation. Thus, the sequence elements that determine the specific points of termination within the tsp region are the elements that account for pausing. Although some of the sequence features that cause RNA polymerase to pause have become apparent, many of the features are still unknown. Also, a large number of different sequences appear to contribute in various ways, thus making it nearly impossible to discern a specific set of sequences. Conceivably, Rho itself could influence the extent of pausing in a terminator region, although there is no evidence for that.
Rho Protein: Structure, RNA Binding, and Helicase Activity.
Rho is a hexamer of a single polypeptide with 419 amino acid residues (67, 68, 201, 203). Electron micrographs reveal that the subunits have a compact globular shape with a diameter of about 4.2 nm and are organized in a flat hexameric ring (16, 96, 194). The small-angle X-ray scattering properties of Rho in solution are consistent with this view (87, 236). Both sequence and functional analyses have shown that the Rho subunit has a distinct domain for binding RNA and another for binding NTPs (15, 57, 58, 59, 60, 61). The fact that Rho can protect 70 continuous residues of poly(C) from digestion with pancreatic RNase suggests that the RNA-binding site in the hexamer consists of a continuous groove that extends across all six subunits (16, 81, 175). However, binding studies with oligo(C) molecules with chain lengths of less than 22 have indicated that only three subunits in the hexamer can bind RNA tightly at one time (88) while three others can bind with a 10-fold-lower binding affinity (267, 268). With oligo(C)10 the association constants, under standard binding conditions, were 4.5 × 106 M–1 for the stronger sites and 0.58 × 106 M–1 for the weaker sites. These results can be explained by interactions between pairs of monomers in the Rho hexamer, in which one subunit of such a dimer exists in a strong conformation and the other exists in a weak conformation (88). The binding of RNA to one of the subunits would stabilize it in the strong conformation and simultaneously stabilize the site in the other subunit of the pair in the weaker conformation. These binding properties are important considerations that have been used in the formulation of models that explain how ATP hydrolysis can be coupled to the change in the binding of RNA to individual subunits and how these changes provide dynamic interactions that can dissociate RNA from transcription complexes.
The RNA-binding domain of Rho has been isolated as an independent entity (179). A fragment consisting of the first 116 amino acid residues binds to short C-rich RNA ligands with nearly the same affinity as do the weaker sites in the intact protein. Since the fragment is monomeric, this result indicates that an intact site for binding a C-rich RNA exists within a single domain and suggests that differences of affinities in the sites in the individual subunits are controlled by protein-protein interaction within the Rho hexamer.
Attempts to identify the residues in the RNA-binding domain that are likely to make direct contact with RNA have focused on a sequence that extends from residue 60 to residue 66 (24; A. Martinez and J. P. Richardson, submitted for publication). This sequence of DGFGFLR bears some resemblance to a conserved sequence motif found in a class of eukaryotic RNA-binding proteins (134, 204). It is also highly conserved in the Rho homolog proteins from evolutionarily diverse bacteria (197). On the basis of the structures of two eukaryotic proteins that have this similar motif and the properties of derivatives with mutational changes in those residues, a model has been proposed that places the Asp and a Gly residue at a β bend and the two Phe and the Arg residues on a β strand, with the side chains extending into the solvent (Martinez and Richardson, submitted). This proposed RNA-binding element uses the Phe residues for stacking interaction with the bases of the RNA, the Arg for critical multivalent, ionic hydrogen bonds, and the Asp residue as a discriminator that reduces the affinity of Rho for non-rut sequences in RNA (A. Martinez, Ph.D. dissertation, Indiana University, Bloomington, 1993).
Rho factor is an RNA-dependent ATPase (163, 165). Its interactions with certain RNA molecules allow it to catalyze hydrolysis of ATP to ADP and Pi in a steady-state reaction. With poly(C), which is a particularly effective activator, the turnover number (k cat) for the reaction is 150 s–1. With λ cro RNA, which is representative of a transcript terminated by Rho action, k cat = 43 s–1 (65).
In the absence of RNA, Rho binds three Mg-ATP molecules per hexamer with high affinity (268) and three more with a much lower affinity (85). Thus, it has the same type of half-of-the-sites reactivity with ATP as with short RNA ligands. Upon adding an RNA activator to the Rho-ATP complex, the three ATP molecules are rapidly hydrolyzed, suggesting that the high-affinity binding site is a true substrate-binding site (257). The overall process of hydrolysis of ATP to ADP and Pi leads to important, but ill-defined, structural changes in the RNA activator (81). Functionally, the interactions between Rho and a nascent RNA that are coupled to ATP hydrolysis lead to dissociation of the RNA from its complex with RNA polymerase and DNA (120, 215, 244). This ability to dissociate a transcription complex is not specific to E. coli RNA polymerase or even a polymerase complex. Rho can apparently terminate transcription of yeast RNA polymerase II when the DNA has a strong rut site (275). It can also dissociate a single-stranded DNA held to the 3' end of an RNA by base-pairing interactions (23). In the latter case, Rho is acting as an RNA-DNA helicase. Although the ability to serve as a helicase may be indicative of its most critical activity, this reaction may also be just a manifestation of a general ability of Rho to displace all components that are noncovalently attached to the 3' end of the RNA. Since the nascent RNA is held to the transcription complex by contacts in the RNA exit site as well as by a limited amount of pairing with the DNA template, the action of Rho on the RNA has to effect dissociation from both contact points. In any case, the actions of Rho on the RNA transcript that are coupled to ATP hydrolysis drive the dissociation of the transcript, thereby causing termination.
Both the ATPase and helicase activities have been useful in revealing the types of interactions that are involved in termination. For instance, the ATPase reaction has helped to identify RNAs that have functional rut sites and has suggested the presence of two functionally distinct sites for RNA in the Rho protein (65, 210, 214, 263). The helicase activity was used to demonstrate that the action of Rho on RNA is undirectional; Rho dissociates DNA molecules base paired at the 3' end of the RNA but not those base paired to residues at the 5' end (23). It has also revealed that Rho is capable of unwinding RNA-RNA helices (25, 253), a property that was very instrumental in showing that a rut site is more than a mere entry point for Rho’s actions. In some carefully documented experiments, Steinmetz and Platt (254) were able to demonstrate that Rho maintained a strong contact with the rut site of a trp t' transcript while it dissociated an oligonucleotide that was base paired to the RNA at a region that was 3' to the rut site. This experimental result is inconsistent with models for Rho action in which the rut site is used merely as an entry point for initiating a linear translocation toward the 3' end of the RNA. Instead, it is more compatible with models in which Rho maintains a contact with the rut region and extends its interactions toward the 3' end either by simultaneously tracking along that part of RNA (the tethered tracking model) or by wrapping the 3' segment of the Rho-bound RNA around the protein with multiple transient contacts (wrapping model).
Models for Rho Action on RNA.
A very appealing model for the directional translocation activity of Rho along an RNA molecule has been proposed by Geiselmann et al. (86). It is based on the apparent symmetry properties of the Rho hexamer (84) and the evidence for the presence of three strong and three weak ATP-binding sites and three strong and three weak RNA-binding sites per hexamer. At the structural level the dimers would be arranged with a pseudodyad symmetry in which one subunit would be in a conformation with a strong RNA-binding site and the other in a slightly different conformation that gives it a weak RNA-binding site. The same deviation in the symmetry would apply to the ATP-binding site, although the strong ATP-binding site need not be on the same subunit of the pair as the strong RNA-binding site. The role of ATP hydrolysis would be to switch the conformations (62, 237). Thus, in a hexameric structure, a switch in conformation of one dimer could cause the 5' segment of an RNA bound to a strong site in one of the subunits in a dimer to now be bound in a weak conformation. Since the other subunit in the dyad symmetric unit would be oriented in the wrong direction to allow that segment of RNA to bind tightly, the 5' segment would dissociate and be replaced by a 3' segment that would have the right orientation to fit into the subunit with the strong site in the switched conformation. A continuation of the process along the other dimers in the hexamer would result in a progressive movement of Rho toward the 3' end of the RNA.
Because the Geiselmann et al. model (86) does not adequately explain the continued contact of Rho with the rut site on an RNA (66, 254). Platt (202) formulated another specific model, also based on the asymmetric dimer, that uses the strong binding sites as the contact points with the rut region and dynamic interactions with the weaker sites to extend the contacts of Rho progressively toward the 3' end. Although this model is compatible with the evidence for tethered tracking, it lacks a simple means for accounting for the directionality of the action of Rho along the RNA.
In an asymmetric dimer, the strong and weak RNA-binding sites could readily be two different conformations of the RNA-binding domain in the two subunits. However, it is less clear whether mere differences in conformation could account for some of the other functional distinctions that have been inferred for different RNA-binding sites in Rho. Although Rho can bind almost as tightly to single-stranded, C-rich DNA as to RNA, DNA molecules do not activate ATP hydrolysis (164). On the basis of studies of activation of Rho-ATPase with mixtures of DNA and RNA, Richardson (210) proposed the existence of two functionally distinct sites: one that will bind either RNA or DNA independently of ATP and another that is specific for RNA and is linked to the binding or hydrolysis of ATP. The fact that poly(dC) readily blocks the binding of oligo(C) to Rho indicates that sites with the strong RNA-binding conformation, at least, also bind DNA (179). Thus, the proposed RNA-specific site is either the RNA-binding domain site in the "weak" conformation or another, unidentified site. One reason for suspecting the existence of another site is the fact that a mutant form of Rho (Rho-1) that is known to be defective in its ATP-dependent RNA interactions but not its ATP-independent RNA interactions is altered in a residue in the ATP-binding domain (291). The mutation is in the part of the ATP-binding domain that is strongly conserved among putative Rho proteins from other microorganisms but is diverged from the ATP-binding domains of other ATPases (197). The uniqueness of that region to the Rho factors could be an indication of a special function such as an ATP-dependent RNA-binding site. However, since no evidence exists that RNA can interact with a site in the ATP-binding domain, the properties of the Rho-1 mutation could be accounted for by an inability to connect ATP hydrolysis properly with interactions at the known site in the RNA-binding domain. In any case, the possibility that Rho has a site for binding RNA outside the known RNA-binding domain needs to be explored further.
Function of Rho in the Cell.
The lack of a stringent sequence requirement for a Rho-dependent terminator raises the possibility that such terminators might be fairly frequent in DNA sequences. In fact, Rho-dependent terminators are commonly found within genes of E. coli, as well as at the ends of genes and operons, their expected locations (212). However, the ones within genes usually only function when the nascent RNA is not being translated by a ribosome (216). Since Rho has to bind to RNA for it to terminate transcription, it will not be able to gain access if the RNA is blocked by a ribosome that is in the process of translating the mRNA as it emerges from the RNA polymerase. This feature of the Rho mechanism thus accounts for the high specificity of its action: terminators will be limited to those sections of the DNA that do not encode functional parts of mRNA, such as at the end of operons. This raises a question of what protects Rho from acting on transcripts that are not translated. This is particularly relevant to the transcription of rRNA. rRNA genes may be naturally devoid of Rho-dependent terminators because the RNA product is highly structured and has very few segments that might serve as a good binding site for Rho. Indeed, Rho has very poor affinity for isolated rRNA molecules (164). In addition, ribosomal proteins bind to the nascent rRNA and this could afford further protection. However, rRNA operons appear to have been devised with a mechanism that gives them even further immunity to Rho action (3, 157, 180). Sequences near the start of the genes establish an antitermination mechanism. How this mechanism operates is not yet known, but one possibility is that it increases the rate of transcriptional elongation by preventing pausing.
NusG Is a Cofactor for Rho Function.
The ability of Rho to function efficiently with certain terminators depends on the presence of another factor called NusG, a 21-kDa protein that was originally isolated as one of the essential components of the λ N-modified antitermination complex (155). A role for NusG in termination was first shown by Sullivan and Gottesman (259), who found that certain Rho-dependent terminators no longer functioned efficiently in cells in which the level of NusG had been depleted. Their work also showed that NusG depletion had no effect on a Rho-independent (intrinsic) terminator and lesser effects on some Rho-dependent terminators than on others. At first this result seemed perplexing because several Rho-dependent terminators had been found to function very efficiently in vitro in purified systems lacking NusG. However, this discrepancy apparently is a consequence of using nonphysiological conditions that overly favor Rho function in vitro (27, 156, 187). Most in vitro transcription systems use low levels of NTPs in order to achieve good incorporation of radioactive labels. However, these conditions cause RNA polymerase to elongate chains at rates of 5 to 10 nt/s instead of the in vivo rate of >40 nt/s. Since the function of Rho is kinetically coupled to the rate of motion of RNA polymerase along DNA, a slow chain growth rate could compensate for defects in Rho function. Indeed, under conditions in which RNA polymerase is able to elongate chains close to the in vivo rate, NusG is required for Rho to cause termination at the first of the intragenic terminators of the lacZ gene (27). Under slower chain growth conditions, Rho has some function by itself at this terminator and that activity is stimulated by NusG. These results thus suggest that NusG is serving to overcome a kinetic deficiency of Rho alone. Nehrke and Platt (187) found that during transcription of trp t' DNA, NusG became associated in a ternary complex with Rho and RNA polymerase. Since NusG has been shown to interact weakly with RNA polymerase core (155) and with Rho (156), it could serve to bring Rho to the transcription complex and facilitate Rho’s interaction with RNA that causes dissociation. Perhaps NusG allows Rho to become associated with the transcription complex shortly after initiation of the RNA chain, thus having it available to "scan" the nascent RNA for a rut sequence.
The Psu Protein of Bacteriophage P4 Inhibits Rho Function.
The Psu protein of satellite bacteriophage P4 is a virion protein (122) that also suppresses Rho-dependent transcription termination in operons of its helper bacteriophage P2 (158). Although Psu works only at Rho-dependent terminators and not at intrinsic terminators, its activity is not operon specific. The expression of Psu in E. coli is sufficient to inhibit Rho-dependent termination in P2 late genes, plasmid operons, and the host chromosome. Therefore, Psu is likely to be a direct inhibitor of Rho factor. Whether Psu inhibits an enzymatic activity of Rho factor or the interaction of Rho with RNA, ATP, NusG, or RNA polymerase is unknown.
A number of distinct mechanisms have evolved to prevent the termination of transcription. Some impinge directly on RNA polymerase and modify the elongating enzyme so that it can processively transcribe through multiple downstream terminators. These processive mechanisms of antitermination are described below in some detail. Other nonprocessive mechanisms have only local effects on termination by RNA polymerase at particular terminators. These terminators are characteristically located near the beginning of an operon and function as transcriptional attenuators. The structures and functions of various attenuators are described in detail in chapter 81 and will be described here only briefly.
Many of the amino acid biosynthetic operons of E. coli and S. typhimurium have transcriptional attenuators that precede the first structural gene of the operon. These attenuators are characteristically regulated by one or more of the amino acids that are the final products of that particular biosynthetic pathway. Availability of the cognate charged tRNA determines whether the ribosome will stall during the synthesis of a leader peptide, and this, in turn, controls which of two mutually exclusive secondary structures is formed in the nascent transcript. Termination is prevented when an antiterminator structure in the nascent RNA, which forms only in the absence of the cognate charged tRNA, prevents the formation of the RNA hairpin which is an essential feature of the intrinsic terminator located just downstream. The most detailed studies of this type of regulation have been done for the trp and his operons.
A somewhat similar situation exists in the tna operon of E. coli, which encodes tryptophanase. The leader region of the tna operon contains a Rho-dependent terminator whose activity is relieved by a mechanism involving the synthesis of a leader peptide when the operon is induced with tryptophan (94, 255).
Suppression of intrinsic termination also occurs in the regulation of the bgl operon (bglG-bglF-bglB) of E. coli, whose protein products are involved in the catabolism of β-glucosides (reviewed in reference 9). In this case, antitermination depends on the BglG protein, which binds as a dimer to specific RNA sequences (8). These RNA sequences overlap the RNA hairpins needed for the functioning of two intrinsic terminators located upstream of the bglG gene and between the bglG and bglF genes (119, 169). Most likely, BglG prevents termination because it interferes with the formation of the terminator hairpins. BglG is converted to an inactive monomer when it is phosphorylated by BglF (7, 8). In the presence of β-glucosides, which are transported into the cell and phosphorylated by BglF, BglF dephosphorylates BglG (7), which then enables BglG to dimerize and bind RNA.
Still another kind of regulated transcriptional attenuator precedes the structural genes of the 11-gene E. coli rpsJ operon, which encodes S10, L4, and other ribosomal proteins (see chapter 90). In this case, the functioning of the intrinsic terminator requires both L4 and NusA (286, 287, 288) and is, therefore, regulated by the concentration of free L4 not yet assembled into ribosomes. RNA polymerase pauses at the attenuator in the absence of additional factors (242), but NusA enhances pausing in a manner that depends on the upper stem-loop structure of the terminator hairpin (241). Stabilization of the paused complex by L4 depends both on NusA and on a hairpin immediately upstream of the terminator hairpin and is prevented when L4 interacts with a specific region in 23S rRNA (289).
There are seven rrn operons in E. coli. When a Rho-dependent terminator is placed downstream from the promoters of the rrnG operon, RNA polymerase is able to transcribe through the terminator and into a downstream reporter gene (3, 4, 26, 116, 180). This type of experiment indicated that an antiterminator element, which could render RNA polymerase insensitive to downstream Rho-dependent terminators, must be linked to the promoters. Subsequent cloning and mutagenesis studies then delineated this element, now known as boxA (17, 157). Identical boxA elements (TGCTCTTTAACA) are located downstream from the P2 promoter in each of the seven rrn operons. Additional boxA elements are also located between the 16S and 23S rRNA genes in each operon. Although boxA is sufficient for antitermination, mutations between P2 and boxA that impair antitermination have also been identified (17). Therefore, antitermination may be regulated by an unknown factor(s) that interacts with this element(s) upstream of boxA.
Analogies between the rrn antitermination system and antitermination by the N protein of bacteriophage λ (see below) suggested that the functional forms of the boxA elements would be RNA, rather than DNA, and that antitermination in the rrn system would also involve the nus gene products that are involved in antitermination by N. Evidence for the involvement of NusB came from experiments showing that chain elongation during the synthesis of rRNA is impaired in an E. coli nusB mutant (243). E. coli strains that lack NusB grow slowly and are cold sensitive (90), possibly in part because of this defect in the synthesis of rRNA.
Antitermination in the rrn system has been achieved in vitro in reactions containing crude E. coli extracts (250). Depletion of NusB from these extracts revealed that NusB is a direct requirement for antitermination and provided evidence for the additional involvement of NusA, NusG, and the nusE gene product, ribosomal protein S10. Since antitermination could not be reconstituted with all of the known Nus factors in the absence of crude extracts, at least one as yet unidentified factor must also be needed for antitermination in the rrn system.
There is a direct and highly specific interaction between NusB and S10, leading to the formation of a heterodimer (172). The NusB-S10 heterodimer then directly binds rrn boxA RNA in vitro (189), consistent with the observation that overproduction of boxA RNA in vivo leads to titration of NusB (78). The excellent correlation between the effects of mutations in rrn boxA on antitermination (17) and their effects on the binding of the NusB-S10 complex to boxA RNA (189) indicates that the recognition of boxA by NusB and S10 is important for antitermination. There are, as yet, no data which directly distinguish whether S10 participates in transcriptional antitermination as a free protein or as a component of the ribosome, but the observation that NusB affects the rate of translation and participates in protein secretion that is coupled to translation (205) indicates that the latter possibility may be correct.
S10 also binds directly and with 1:1 stoichiometry to RNA polymerase (171). An important consequence of the interaction of S10 with RNA polymerase, NusB, and boxA RNA would be the formation of a loop in the nascent transcript, in analogy with the situation that occurs in transcription complexes modified by the N protein of bacteriophage λ (109; see Fig. 4A). When transcription complexes synthesizing rRNA in vitro are isolated and analyzed, they contain NusB and NusG, whose stable association depends on the presence of a boxA element in the DNA template (155). Therefore, antitermination in this system probably depends on the stable association with RNA polymerase of a ribonucleoprotein complex containing boxA RNA, several Nus proteins, and at least one unidentified factor.
The rrn boxA elements enable RNA polymerase to transcribe through Rho-dependent terminators but not through intrinsic terminators like those at the ends of the rrn operons (4). The presence of stably associated NusG in elongation complexes synthesizing rRNA (155) may indicate how this occurs. Since an interaction of Rho with NusG (156) is important for termination at many Rho-dependent terminators (156, 187, 259), one possibility is that the presence of other Nus factors on RNA polymerase in the vicinity of NusG may prevent Rho from interacting with NusG. Alternatively, the stably associated NusG molecule may sequester the Rho factor that is bound to the nascent transcript and prevent it from moving in a 5'-to-3' direction toward the transcription bubble (156).
Transcriptional regulation during the lytic growth of bacteriophage λ involves the λ N and Q proteins in a temporally controlled cascade of two antitermination mechanisms (47,109, 221; Fig. 3). After infection of E. coli cells by λ, the host RNA polymerase initiates transcription at the early λ promoters pL and pR. It then terminates transcription at the Rho-dependent terminators tL1 and tR1 after synthesizing mRNAs for the Cro and N proteins. When N is made, it prevents termination at all of the terminators in both early operons (i.e., tL1 to tL4 and tR1 to tR3). This allows RNA polymerase to more efficiently transcribe the cII gene, which is needed to establish lysogeny, the O and P genes, which are needed for λ DNA replication, the red, gam, bet, int, and xis genes, which are involved in vegetative and integrative recombination, and the Q gene, which encodes the key regulator of late transcription. Q protein, in turn, prevents termination by RNA polymerase molecules which initiate at the late promote p' R. This enables RNA polymerase to transcribe through t' R and, potentially, other terminators and synthesize mRNAs for all of the late genes involved in host cell lysis and bacteriophage morphogenesis. In the alternative pathway leading to lysogeny, CII protein activates transcription of the cI and int genes from the pRE and pI promoters. This leads to production of Int protein, which integrates the λ DNA into the E. coli chromosome, and CI protein, which represses further initiation at pL and pR and activates transcription from its own promoter, pM.
The N gene requirement for RNA polymerase to transcribe λ DNA downstream from tL1 and tR1 led Roberts to postulate that N might be an antitermination factor (220). This hypothesis was first supported by an experiment showing that RNA polymerase must initiate at pL upstream from tL1 in order for N to stimulate transcription downstream from tL1 (167). Similarly, an experiment done in vitro revealed that RNA polymerase must initiate at pR in order for N to stimulate transcription of the downstream R gene (100). Confirmation that N is truly an antitermination factor came with the demonstration that the RNA across tL1 is continuous in the presence of N (166).
Like most other transcriptional regulatory mechanisms, antitermination by N requires a specific cis-acting sequence, the nut site. A nut site is located between the promoter and the first terminator of each early operon. This kind of arrangement is also found in the related lambdoid bacteriophages, φ21 and P22, which also have N genes and nut sites (73). It was as a consequence of experiments with hybrid bacteriophages, in which a DNA segment from bacteriophage φ21 or P22 replaced λ DNA in the immunity region between the N and cro genes, that it was first realized that the functioning of N at downstream terminators must involve specific promoter-proximal sequences (80). In these experiments, it was found that the N proteins of λ and φ21 can only function at the downstream terminators of the pR operon if the promoters and their proximal downstream sequences were derived from the cognate bacteriophage. This implied that N of λ can only function with its own nut site and not with that utilized by N of bacteriophage φ21. Furthermore, N must modify RNA polymerase near the promoter in such a way that it remains modified when it arrives at a terminator that can be quite far downstream. Similar conclusions were drawn from experiments fusing the λ pL promoter to E. coli gal and trp operons that contained polar nonsense mutations (2, 72, 235). Polarity on downstream gene expression is the consequence of premature termination of transcription induced by Rho factor when a nonsense mutation prevents ribosomes from translating the nascent mRNA (212). N suppressed the polarity caused by Rho factor and stimulated the expression of downstream gal and trp genes only if transcription initiated at the λ promoter and not if it initiated at the gal or trp promoter. In effect, N was acting at a specific sequence near λ pL to convert RNA polymerase into a "juggernaut" that could transcribe through downstream terminators.
The identification and sequencing of mutations in nutL that impaired N-dependent transcription downstream from tL1 led to a description of the nutL site (228) and then to identification of the virtually identical nutR site (224) (Fig. 3). When such a nut site was inserted between a promoter and a terminator, N was able to activate the expression of a reporter gene located downstream from the terminator (52, 200). Therefore, a nut site is both necessary and sufficient to make an operon a target for antitermination by N. In this kind of experiment, N enables RNA polymerase to transcribe through both intrinsic and Rho-dependent terminators. Consistent with this, some of the terminators in the λ early operons are Rho-dependent (e.g., tL1 and tR1 [220]) and some are intrinsic (e.g., tR2 [41, 233]).
The nut sites of λ, φ21, and P22 are characterized by two kinds of genetically defined elements, boxA and boxB, and both are important for antitermination in at least some circumstances (56, 77, 195, 196, 228). The boxA elements of these bacteriophages are very similar (73), and their consensus sequence (CGCTCTTTA) is nearly identical to the boxA sequences of the E. coli rrn operons (TGCTCTTTA). The boxB elements vary considerably in sequence from one bacteriophage to another, but all have hyphenated dyad symmetry in the DNA, allowing the formation of a stem-loop structure in the nascent RNA. As demonstrated in genetic experiments using chimeric N proteins, these hairpins are differentially recognized by the arginine-rich amino-terminal regions of Nλ, N21, and NP22 (150). Arginine-rich RNA-binding motifs are found in many RNA-binding proteins, including the human immunodeficiency virus Tat protein, an antiterminator of human immunodeficiency virus transcription in infected human cells (44).
The selection of E. coli mutants that were resistant to λ infection or to a killing effect of N led to the identification of mutant strains in which N cannot function. Some of these mutations (e.g., ron, groN785, and nusC60) are in the β subunit of RNA polymerase and create a form of RNA polymerase that cannot be modified by N (89, 92). The nusD mutations are in the rho gene and create forms of Rho factor that are resistant to antitermination by N, although N can still prevent termination at intrinsic terminators in such strains (49). Most important, this genetic analysis identified the nusA, nusB, and nusE genes, all of which encode proteins that participate in antitermination by N (74, 75, 79, 135). Surprisingly, nusE encodes ribosomal protein S10 (79). The development of in vitro systems containing crude E. coli extracts for analyzing antitermination by N (95, 100, 124) made possible the purification of N (107). This also led to in vitro complementation assays showing that NusA, NusB, and S10 all participate directly in antitermination by N (48, 50, 91, 95, 118, 261).
An important goal was the reconstitution of antitermination by N in vitro in reactions containing purified proteins. In the presence of NusA, NusB, S10, and Rho, the stimulation by N of λ transcription in vitro is very weak (155). Purification of an activity in an E. coli extract that enabled N to strongly stimulate transcription in the presence of the other factors led to the identification of NusG (155). Simultaneously, the nusG gene was identified on the basis of a mutation that suppresses the effect of the nusA1 mutation on antitermination by N (260). Therefore, the four E. coli proteins known to be involved in antitermination by N are NusA, NusB, NusG, and S10. N enables RNA polymerase to processively transcribe in vitro through all of the intrinsic and Rho-dependent terminators of the λ early operons in the presence of these factors (53, 173).
A characteristic feature of processive antitermination mechanisms is that RNA polymerase must somehow maintain a termination-resistant form far downstream from the antiterminator sequence. For antitermination by N, there is now considerable evidence that an RNA loop initially connects the λ nut site to the RNA polymerase (Fig. 4A), so that RNA polymerase is still associated with the nut site when it arrives at a downstream terminator.
The first evidence that the nut site is made of RNA, rather than DNA, as predicted by this model (103), came from an experiment showing that a frameshift mutation enabling ribosomes to translate beyond the end of the cro gene and into the boxA element of the λ nutR site prevents antitermination by N (195). Similarly, N cannot function when the entire nut site is made to be part of a translated region (270, 292). The observation that the nut site in the nascent RNA is protected from chemical modification or RNase attack when transcription in vitro is carried out in the presence of N, NusA, NusB, NusG, and S10 provided direct evidence that the nut site is made of RNA (188). This N-dependent protection of the nut site RNA is specific because it does not occur when there is a mutation in boxB or NusA that prevents antitermination by N. The interaction of N with the elongation complex is RNA dependent because N is released when elongation complexes are digested with large amounts of RNase T1 (118).
Another prediction of the model shown in Fig. 4A is that N and the Nus factors remain stably associated with RNA polymerase during chain elongation. When gel filtration was used to isolate elongation complexes from reactions containing either crude E. coli extracts or purified N, RNA polymerase, and the Nus factors, the N-modified elongation complexes were found to contain near-stoichiometric amounts of N and all four Nus factors (118, 171). None of these factors is simply a DNA-binding protein, because the association of all of the factors with the DNA template depends on RNA polymerase. As well, the association with the template of all but S10 depends on the presence of a nascent transcript. Although NusA and S10 can bind to RNA polymerase in a sequence-independent manner (105, 171), the stable association of N, NusB, and NusG with the elongation complex depends on the presence of a wild-type nut site (171). In the case of N, its interaction with the elongation complex was also demonstrated by coimmunoprecipitation of N and the nascent RNA with antibody against N (13). In these experiments it was also made clear that N becomes associated with the elongation complex only after RNA polymerase has transcribed past the nut site.
In principle, N and the Nus factors could bind to the nut site in a nascent transcript and still not use RNA looping to make direct contact with RNA polymerase. For two reasons, however, this is unlikely to be true: first, the ron and groN785 mutations in the RNA polymerase β subunit prevent the stable binding of N, NusB, and NusG to the elongation complex, even when the nut site is the wild type (171); and second, protection of the nut site from RNase is incomplete when RNA polymerase has these same mutations (188). These observations indicate that the surface of RNA polymerase must properly align the elongation factors so that they can stably bind the nut site RNA, and they are only easy to understand in the context of an RNA looping model (Fig. 4A). Consistent with the RNA looping model and the unimportance in this model of the nut site DNA, N can engage the elongating RNA polymerase after it has passed the nut site, even when the nut site DNA has been removed from the template by digestion with a restriction enzyme (271).
For both the λ nut sites and the rrn boxA elements, the critical antiterminator elements and their bound proteins would initially associate with RNA polymerase because the nascent RNA serves as a "tether" that creates a higher effective concentration of the antiterminator element in the vicinity of RNA polymerase (103, 188, 271). Once a stable complex containing all of the appropriate factors forms on the surface of RNA polymerase, the RNA loop is probably dispensable. Indeed, the primary transcripts of the rrn operons are eventually processed to form mature rRNAs and tRNAs, and the pL RNA of λ is cleaved between nutL and tL1 (166, 252).
N is the only elongation factor which can directly bind the nut site RNA in a gel mobility shift assay (37; J. Mogridge, T.-F. Mah, and J. Greenblatt, submitted for publication), and this interaction requires only the boxB hairpin. Interaction of NusA, NusB, NusG, and S10 with the nut site RNA then depends on both protein-protein and protein-RNA interactions (Mogridge et al., submitted). As mentioned above, the NusB-S10 heterodimer does bind to rrn boxA RNA in a gel mobility shift assay, but its binding to λ boxA cannot be detected in the same kind of assay (189; Mogridge et al., submitted). Two of the three nucleotide changes which distinguish rrn boxA from λ boxA are each sufficient to prevent the binding of NusB and S10. Consistent with this, λ boxA cannot function as an antiterminator in the absence of boxB and N (52, 228). Only when N and all four Nus factors are present is the λ boxA sequence in the nascent transcript protected from RNase during transcription in vitro (188). The λ boxA element thus behaves as a defective form of rrn boxA whose ability to bind the NusB-S10 complex and induce antitermination has come to depend on N and boxB.
Many of the protein-protein interactions involved in this control system (Fig. 4B) were first identified by protein affinity chromatography. The use of crude E. coli extracts in these experiments ensured that the interactions that were identified are highly specific. For example, the only protein in a crude E. coli postribosomal supernatant fraction that is bound by a column containing immobilized S10 is NusB (172). This interaction is very weak, with a Kd of approximately 10–6 M, and seems to be entirely nonionic, since it is not dissociated with 1 M salt. Similarly, the only protein in a crude E. coli extract that is reproducibly bound by a column containing immobilized N is NusA (104). Again, the interaction is weak, with a Kd of about 10–7 M, but in this case it is at least partly ionic. Further confirmation that these interactions are functionally important has come from gel mobility shift experiments. First, NusB and S10 bind cooperatively in vitro to the rrn boxA element (189). Second, while NusA does not bind by itself to rrn boxA or the λ nut site (189), NusA does bind to an N-nut site complex (Mogridge et al., submitted).
Mutations in all five nucleotides of the loop of boxB (Fig. 3) impair antitermination by N (37, 56). Interestingly, only mutations at loop positions 1, 3, and possibly 5 reduce or impair the binding of N (37; Mogridge et al., submitted), while mutations at positions 2 and 4 do not affect the binding of N and, instead, impair the binding of NusA (Mogridge et al., submitted). This implies that NusA interacts with RNA as well as with N and, indeed, the nusA1 mutation affects NusA’s interaction with the nut site RNA, rather than its interaction with N (104; Mogridge et al., submitted). Similar gel mobility shift experiments with boxA mutants have revealed that boxA is also important for the binding of NusA to the N-nut site complex (Mogridge et al., submitted). This may imply that the nut site assumes a folded structure, perhaps a pseudoknot, in which boxA is brought into the vicinity of the loop of boxB in order to create a more complex structure that is recognized by NusA.
Three of the Nus factors, NusA, S10, and NusG, interact directly with RNA polymerase. For example, an affinity column containing immobilized NusA binds only the core component of RNA polymerase (Kd ≈ 5 × 10–8 M) from a crude E. coli extract derived from uninfected cells (105). Consistent with this, NusA can directly influence pausing by RNA polymerase and termination at intrinsic terminators (see above). The ability of NusA to bind N and RNA polymerase implies that NusA is an adaptor that couples N to RNA polymerase (102). If that is so, however, what is the role of the nut site RNA? Cross-linking experiments with NusA that is derivatized with a photoactivatable cross-linking reagent have revealed that N binds about 5 to 10 times as tightly to free NusA as it does to a NusA-RNA polymerase complex (J. Li and J. Greenblatt, unpublished data). The tethering effect of the RNA during chain elongation should then overcome this natural inclination of N to bind free NusA and ensure that it interacts with the NusA on elongating RNA polymerase. Indeed, as predicted by this model, providing an elevated concentration of N overcomes the sequence specificity for antitermination that is normally imposed by the nut site (230).
NusG and S10 also interact with RNA polymerase (Fig. 4B). In the case of NusG, this was demonstrated in an affinity chromatography experiment in which immobilized RNA polymerase bound the NusG, as well as the NusA, from a crude E. coli extract (155). Consistent with this, NusG can accelerate elongation by RNA polymerase (28) and reduce the efficiency of an intrinsic terminator in vitro (159) and, of course, it is a cofactor for termination by Rho (see above). Gel filtration experiments have shown that S10 forms a 1:1 complex with DNA-bound RNA polymerase (171). Unlike the situation with NusA, S10 can bind to RNA polymerase in the presence or absence of σ 70 (171). Despite this, however, no effect of S10 alone on initiation or elongation has yet been described.
Once N and NusA are bound to the nut site RNA, a gel mobility shift assay can be used to demonstrate a specific interaction of this complex with the core component of RNA polymerase (37; Mogridge et al., submitted). The formation of this quaternary complex is prevented by mutations in boxA and in the loop of boxB (37; Mogridge et al., submitted). Moreover, only this quaternary complex can be bound and supershifted by NusB, S10, and NusG in the same kind of assay (Mogridge et al., submitted). Since the interaction of NusG with this complex facilitates the subsequent binding of NusB in the absence of S10, there is likely to be a direct interaction between NusG and NusB (Fig. 4B).
Studying the effects of mutations that impair antitermination on the formation of these complexes has also been revealing (Mogridge et al., submitted). For example, certain mutations in λ boxA that do not affect the binding of NusA to the N-nut site complex do impair the binding of NusG, as well as NusB and S10. Therefore, NusG may also interact with boxA in the context of a λ or rrn elongation complex. The ron and groN785 mutations in the RNA polymerase β subunit allow the formation of a quaternary complex containing N, NusA, and RNA polymerase but prevent the subsequent association of NusG. Therefore, sequencing of these mutations may help define the binding site for NusG on RNA polymerase.
Conversion of λ boxA (CGCTCTTAC) to consensus boxA (CGCTCTTTA) in the context of a λ nut site suppresses the effects of the nusA1 and nusE71 mutations on antitermination by N (77). In gel mobility shift assays, however, the use of consensus boxA in a λ nut site does not facilitate the binding of NusA or S10 (Mogridge et al., submitted). Instead, it enables NusB to bind strongly to a quaternary complex containing N, NusA, and RNA polymerase in the absence of S10 or NusG. Therefore, it is almost certainly NusB, rather than S10, that directly recognizes rrn boxA or consensus boxA. There is also genetic evidence that NusB may interact with N (269). If this is so, the interaction of N with NusA (104) may then explain why facilitated binding of NusB overcomes the assembly defect imposed by the nusA1 mutation.
The protein-protein and protein-RNA interactions that have been identified in this system are summarized in Fig. 4B. Surprisingly, N, all four Nus factors, and the nut site RNA can be assembled into a ribonucleoprotein complex that is stabilized by and associates with the core component of RNA polymerase in the complete absence of a DNA template (Mogridge et al., submitted). A remarkable feature of this interactive network is that each element of the system interacts with at least two, and more often three or four, other elements of the system. Moreover, the interactions that have been identified in direct binding experiments are typically quite weak, and the interactions that have been identified only in genetic experiments or by supershifts in gel mobility shift assays are presumably even weaker. This almost certainly explains why assembly of the complete complex is a highly cooperative process (118, 171, 188; Mogridge et al., submitted). The use of multiple weak interactions is likely to be a common feature in complicated macromolecular complexes that are assembled on a nucleic acid scaffold and must periodically be disassembled and reassembled to regulate various biological processes (1, 117, 256, 285).
When a terminator is located within a few hundred base pairs downstream of a nut site, N is able to prevent termination in vitro at both intrinsic and Rho-dependent terminators with NusA as the only host bacterial factor present in the reaction (173, 272). This is probably also true in vivo since N can prevent termination at tR1 just downstream from nutR in the rightward early operon when boxA is deleted from the nutR site (199, 200, 292). Under these conditions it is likely that NusB, NusG, and S10 cannot participate in antitermination because their stable association with the elongation complex depends on boxA (118; Mogridge et al., submitted). Moreover, in the absence of boxA, antitermination at tR1 is no longer affected by the nusB5 mutation and is less affected by the nusE71 mutation (199).
What then are the roles in antitermination of NusB, NusG, and S10? These proteins are all important for the functioning of N in the context of λ growth (74, 75, 79, 135, 260). In vivo, antitermination is highly processive, and N can prevent termination at sites located as much as 5 to 10 kb downstream from nutL or nutR (Fig. 3). In vitro, the antitermination by N is also highly processive provided that NusA, NusB, NusG, and S10 are all present in the reaction (53, 173). Only under these conditions are N and all four Nus factors stably bound to the elongating transcription complex (118, 171). In contrast, the association of N with the elongation complex is unstable when NusA is the only host factor in the reaction. In the absence of the stabilizing influences of NusB, NusG, and S10, the dissociation of N should become more and more favored as the size of the tethering RNA loop grows and its localizing influence diminishes (188) (Fig. 4A). Therefore, NusB, NusG, and S10 are stability factors that increase the processivity of a core antitermination complex containing N and NusA.
The most important mechanism for antitermination by N must lie in the unstable core antitermination complex containing only N and NusA (173, 272). Under these conditions, N inhibits pausing by RNA polymerase in a manner that depends on both NusA and the presence of a nut site upstream from the pause sites (173). In view of observations that mutations in RNA polymerase that alter its elongation rate through pause sites have parallel effects on termination at both intrinsic and Rho-dependent terminators (125, 127, 128, 143, 174), this effect of N on the rate of chain elongation seems very likely to explain its ability to prevent termination.
In the core antitermination complex, elongation by RNA polymerase could potentially be influenced by three elements: N, NusA, or the boxB RNA hairpin of the nut site. It seems unlikely that boxB binds specifically to RNA polymerase because the mutations in the loop of boxB that impair antitermination (37, 56, 228) affect the binding to boxB of either N or NusA (37; Mogridge et al., submitted). It is also unlikely that NusA prevents termination because, on its own, NusA enhances pausing by RNA polymerase (64, 69, 133, 136, 148, 231) and increases the efficiencies in vitro of some intrinsic terminators (99, 108, 233). Therefore, the role of NusA may simply be to increase the association between the N-nut site complex and RNA polymerase and to bind N and RNA polymerase in such a way as to correctly position N at its RNA polymerase-binding site. Recent experiments have indicated that elevated concentrations of N can prevent termination in the absence of NusA (53). If this observation is a reflection of the usual mechanism by which N prevents termination, then N itself, which may only adopt the appropriate conformation when it binds to boxB, must interact directly with a key site on RNA polymerase. This site on RNA polymerase has yet to be defined by mutations since the ron and groN785 (89, 92) mutations affect the binding of NusG, rather than the binding of N (Mogridge et al., submitted).
How then might an interaction of N with RNA polymerase prevent pausing and termination of transcription in the context of our current models for the regulation of chain elongation (see above and Fig. 2)? The formation of a hairpin structure in the nascent transcript is important for pausing by RNA polymerase at many sites. Moreover, NusA can be cross-linked to the nascent RNA (162) and, in the case of the pause site of the trp attenuator, there is evidence that NusA is in the vicinity of the hairpin and stabilizes its formation (146). Similar hairpins are also key features of all intrinsic terminators and some Rho-dependent terminators. One possibility is that N suppresses pausing and termination because N, normally in conjunction with boxB, occupies part of the physical space that pause and terminator hairpins would otherwise have to occupy in the RNA exit channel on RNA polymerase. An alternative possibility derives from observations that RNA polymerase enters a discontinuous mode of elongation about 8 nt upstream from the pause site in the his attenuator (266) and about 9 nt upstream from the release point in the intrinsic λ terminator tR2 (191). If this mode of elongation is critical for pausing and termination, contact of N with RNA polymerase may somehow keep the leading edge of RNA poly-merase moving forward in a continuous manner. Interestingly, when RNA polymerase is in the discontinuous mode of elongation, the 3' end of the nascent RNA is a particularly effective substrate for cleavage by GreB (19, 20, 190). Therefore, if entry into the discontinuous mode of elongation is a stochastic process, it is also possible that N suppresses pausing and termination because, like the Gre factors, it causes cleavage of the nascent RNA as soon as the leading edge of RNA polymerase is arrested and strain begins to build in the nascent transcript. This would eventually allow RNA polymerase to pass through the pause site or terminator without ever forming a fully strained transcription complex.
Antitermination by N has a role in the timing of λ gene expression and could be regulated (Fig. 3). As well, the effectiveness of antitermination by N could affect the relative probabilities that λ will enter the lytic cycle or, instead, opt for lysogeny. In the latter case, N quickly disappears because it is degraded by the Lon protease (97, 131) and has a half-life of only 2 to 5 min (101, 137, 234). The involvement of so many bacterial proteins in antitermination by N provides another opportunity for regulation. Indeed, the Nus factors also control elongation during the transcription of the rrn operons (243, 250), and these elongation rates are regulated as a function of cell growth rate by an unknown mechanism (264). Such a control mechanism could operate on rrn boxA or on a predicted hairpin structure in the RNA upstream from rrn boxA (17, 157).
An additional control mechanism seems to operate on λ boxA. The boxA element in the λ nutR site is apparently recognized by an unknown cellular repressor of antitermination because NusB is needed for antitermination at tR1 only when wild-type boxA is present (199). Therefore, boxA and NusB have dual roles in λ transcription: they make antitermination more processive by stabilizing the association of N with the elongation complex (118, 171; Mogridge et al., submitted); and they interfere with a cellular mechanism that can inhibit antitermination.
Antitermination by N is also prevented by the nun gene product of the lambdoid bacteriophage HK022, and this prevents the growth of λ on an HK022 lysogen (218). Like N, Nun contains an arginine-rich motif (193), and it probably competes with N for binding to boxB (14, 121, 222). Recognition of the nut site by Nun, however, leads to premature termination rather than antitermination (121, 222, 247). Remarkably, even though the effects of N and Nun are opposite, regulation by Nun is affected by mutations in nusA, nusB, and nusE and requires NusG (218, 219, 259). Moreover, a mutation in boxA can make Nun behave as an antitermination factor rather than a termination factor (222). It is not yet understood how N and Nun can functionally interact with the same set of genetic determinants and yet have completely opposite effects on termination.
Nun has no effect on the growth of HK022 nor any known effect on HK022 transcription. HK022 does have antiterminator sequences in its early operons, but the functioning of these sequences does not depend on any HK022-encoded N-like protein (192). In fact, the HK022 nut-like sequence enables RNA polymerase to transcribe through an intrinsic terminator in vitro in the absence of other factors (42). The identification of mutations in a putative zinc-binding region of the β' subunit of RNA polymerase that prevent the HK022 antiterminator from functioning suggests that the antiterminator DNA or RNA uses DNA or RNA looping to maintain a factor-independent interaction with RNA polymerase during elongation in order to prevent termination (42).
The Q protein of bacteriophage λ prevents termination at t' R, an intrinsic terminator which lies 194 bp downstream from the λ late promoter p' R (98, 221). Q also prevents termination at other intrinsic and Rho-dependent terminators which are placed downstream from p' R (71, 279, 280). Therefore, like N, it recognizes a signal near the promoter and modifies the RNA polymerase so that it will become and remain termination resistant as it moves further downstream. Efficient transcription of all of the genes in the 26.5-kb λ late operon depends on Q.
Deletion mapping and the construction of chimeric promoters revealed that the qut site recognized by Q spans the initiation site of the p' R promoter, with important elements both upstream and downstream from +1 (Fig. 3) (130, 248, 279, 282). RNA polymerase molecules which initiate at p' R pause for several minutes at +16 or +17 before continuing on and terminating at t' R (99, 130). The transcribed region of qut downstream from +1 acts as the pause signal, as point mutations at +2 and +6 prevent both pausing at +16 and antitermination by Q (282). These nucleotides are in the DNA region which is melted when RNA polymerase pauses at +16 (130), and the use of hybrid templates containing mutations in only one DNA strand has shown that the sequence of the nontemplate strand is mainly responsible for causing RNA polymerase to pause (217). Detailed mutagenesis of this sequence has indicated that it causes pausing by RNA polymerase to the extent that it resembles the –10 region of a canonical σ 70 promoter (J. W. Roberts, personal communication). Although σ 70 is normally released by the time that RNA polymerase arrives at +16 (111), perhaps the interaction with a melted –10 region of RNA polymerase itself or σ 70 that had still not been released would cause RNA polymerase to enter an inchworming cycle and lead to a long kinetic pause.
Unlike N, which binds to RNA, Q is a site-specific DNA-binding protein (282). Mutations at –13 and –15 within the p' R promoter prevent the binding of Q to the DNA between –10 and –30 and block antitermination. Although Q can bind to this sequence in the absence of RNA polymerase, it binds to the same region when RNA polymerase is stalled at +16. The mutation at +2 which abolishes Q activity does so even if the RNA polymerase is stalled at +16 by nucleotide deprivation (282). This suggests that the nucleotides downstream from +1 not only stall the polymerase but also cause it to adopt a special conformation, allowing modification by Q. If NusA is present, Q then induces changes in the contacts between RNA polymerase and the promoter, especially between –2 and +6. It is possible that these changes are maintained during chain elongation and are important for the action of Q.
The binding of Q to the qut site when RNA polymerase has paused at +16 accelerates the polymerase out of the pause site and through t' R into the late gene region(99). Q can also prevent pausing at the pause sites of the λ Rho-dependent terminator tR1 when that terminator is artificially fused to p' R (280). As in the case of N, this ability of Q to inhibit pausing may explain its ability to suppress termination at both intrinsic and Rho-dependent terminators. This type of regulation may also be quite universal: in eukaryotic cells, both cellular and viral activator proteins that bind DNA or RNA can often inhibit pausing or premature termination by RNA polymerase II (109).
Transcriptional antitermination in vitro by the Q protein of bacteriophage 82 does not require NusA or any other bacterial host factor (278). Moreover, the λ Q protein can act alone on the stalled poly-merase at +16, although the addition of NusA greatly facilitates antitermination in this case (99). There is also no indication that additional host factors are needed to maintain the processivity of Q-mediated antitermination. Therefore, the antitermination activity of Q may result from a direct contact between Q and RNA polymerase that occurs when the polymerase is stalled at +16 and Q is appropriately positioned on the qut site. Since Q and N both inhibit pausing by RNA polymerase, it is even conceivable that they interact with the same site on the RNA polymerase.
How Q prevents termination at sites far downstream from p' R is unknown. For antitermination proteins that bind to RNA, the association of the control signal and antiterminator protein with the transcription complex can be maintained by RNA looping, as shown in Fig. 4A. For a DNA-binding antiterminator protein like Q, the analogous possibility would be DNA looping (109; Fig. 5A). Another possibility is that Q is released from the DNA and travels with the RNA polymerase when it leaves the pause site at +16 (Fig. 5B). According to this model, since Q cannot engage the polymerase without a qut site, it would have to trap a particular conformation of the RNA polymerase induced by the qut site or Q at +16. This would then force Q to remain bound to the polymerase during elongation. Experiments that distinguish between these models will be crucial for understanding how Q acts.
Research during the last several years has led to a much better understanding of the complexities of the enzymology and regulation of chain elongation by bacterial RNA polymerase. Most notably, the static picture of the transcription bubble that prevailed only a few years ago (276) has given way to our current view that the elongation complex changes dynamically in response to specific DNA sequences encountered by the transcribing RNA polymerase (138, 139, 190). Much of this progress has been a consequence of the development of methods to analyze elongation complexes arrested at specific base pairs on the DNA template (145, 154, 190). Progress should continue to be rapid in the years to come as continuing genetic and structural studies on RNA polymerase are combined with more detailed experimentation on the many mechanisms that control pausing by RNA polymerase and termination at intrinsic and Rho-dependent terminators.
References
1. Cell 80:859–868.
2. Adhya, S., M. Gottesman, and B. de Crombrugghe. 1974. Release of polarity in E. coli by gene N of phage λ: termination and antitermination of transcription. Proc. Natl. Acad. Sci. USA 71:2534–2538.
3. Aksoy, S., C. L. Squires, and C. Squires. 1984. Evidence for antitermination in Escherichia coli rRNA transcription. J. Bacteriol. 159:260–264.
4. Albrechsten, B., C. L. Squires, S. Li, and C. Squires. 1990. Antitermination of characterized transcriptional terminators by the Escherichia coli rrn leader region. J. Mol. Biol. 213:123–134.
5. Alifano, P., F. Rivellini, D. Limauro, C. B. Bruni, and S. Carlomagno. 1991. A consensus motif common to all rho-dependent prokaryotic transcription terminators. Cell 64:553–563.
6. Altmann, C. R., D. E. Solow-Cordero, and M. J. Chamberlin. 1994. RNA cleavage and chain elongation by Escherichia coli DNA-dependent RNA polymerase in a binary enzyme RNA complex. Proc. Natl. Acad. Sci. USA 91:3784–3788.
7. Amster-Choder, O., F. Houman, and A. Wright. 1989. Protein phosphorylation regulates transcription of the β-glucoside utilization operon in E. coli. Cell 58:847–855.
8. Amster-Choder, O., and A. Wright. 1992. Modulation of the dimerization of a transcriptional antiterminator protein by phosphorylation. Science 257:1395–1398.
9. Amster-Choder, O., and A. Wright. 1993. Transcriptional regulation of the bgl operon of Escherichia coli involves phosphotransferase system-mediated phosphorylation of a transcriptional antiterminator. J. Cell. Biochem. 51:83–90.
10. Andrews, C., and J. P. Richardson. 1985. Transcription termination factor rho mediates simultaneous release of RNA transcripts and DNA template from complexes with Escherichia coli RNA polymerase. J. Biol. Chem. 260:5826–5831.
11. Arndt, K., and M. J. Chamberlin. 1990. RNA chain elongation by Escherichia coli RNA polymerase. Factors affecting the stability of elongating ternary complexes. J. Mol. Biol. 211:79–108.
12. Arndt, K. M., and M. J. Chamberlin. 1988. Transcription termination in Escherichia coli: measurement of the rate of enzyme release from rho-dependent terminators. J. Mol. Biol. 202:271–285.
13. Barik, S., W. Ghosh, W. Whalen, D. Lazinski, and A. Das. 1987. An antitermination protein engages the elongating transcription apparatus at a promoter proximal recognition site. Cell 50:885–899.
14. Baron, J., and R. A. Weisberg. 1992. Mutations of the phage λnutL region that prevent the action of Nun, a site-specific transcription termination factor. J. Bacteriol. 174:1983–1989.
15. Bear, D. G., C. L. Andrews, J. D. Singer, W. D. Morgan, R. A. Grant, P. H. von Hippel, and T. Platt. 1985. Escherichia coli transcription termination factor rho has a two-domain structure in its activated form. Proc. Natl. Acad. Sci. USA 82:1911–1915.
16. Bear, D. G., P. S. Hicks, K. W. Escudero, C. L. Andrews, J. A. McSwiggen, and P. H. von Hippel. 1988. Interactions of Escherichia coli transcription termination factor rho with RNA. II. Electron microscopy and nuclease protection experiments. J. Mol. Biol. 199:623–635.
17. Berg, K. L., C. Squires, and C. L. Squires. 1989. Ribosomal RNA operon antitermination. Function of leader and spacer region boxB-boxA sequences and their conservation in diverse micro-organisms. J. Mol. Biol. 209:345–358.
18. Borukhov, S., J. Lee, and A. Goldfarb. 1991. Mapping of a contact for the RNA 3' terminus in the largest subunit of RNA polymerase. J. Biol. Chem. 266:23932–23935.
19. Borukhov, S., A. Polyakov, V. Nikiforov, and A. Goldfarb. 1992. GreA protein: a transcription elongation factor from Escherichia coli. Proc. Natl. Acad. Sci. USA 89:8899–8902.
20. Borukhov, S., V. Sagitov, and A. Goldfarb. 1993. Transcript cleavage factors from E. coli. Cell 72:459–466.
21. Bremer, H., and P. P. Dennis. 1987. Modulation of chemical composition and other parameters of the cell by growth rate, p. 1527–1542. In F. C. Neidhardt, J. C. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, vol. 2. American Society for Microbiology, Washington, D.C.
22. Brendel, V., G. H. Hamm, and E. N. Trifonov. 1986. Terminators of transcription with RNA polymerase from Escherichia coli: what they look like and how to find them. J. Biomol. Struct. Dyn. 3:705–723.
23. Brennan, C. A., A. J. Dombroski, and T. Platt. 1987. Transcription termination factor rho is an RNA-DNA helicase. Cell 48:945–952.
24. Brennan, C. A., and T. Platt. 1991. Mutations in an RNP1 consensus sequence of rho protein reduce RNA binding affinity but facilitate helicase turnover. J. Biol. Chem. 266:17296–17305.
25. Brennan, C. A., E. J. Steinmetz, P. Spear, and T. Platt. 1990. Specificity and efficiency of rho-factor helicase activity depends on magnesium concentration and energy coupling to NTP hydrolysis. J. Biol. Chem. 265:5440–5447.
26. Brewster, J., and E. A. Morgan. 1981. Tn9 and IS1 inserts in a ribosomal ribonucleic acid operon of Escherichia coli are incompletely polar. J. Bacteriol. 148:897–903.
27. Burns, C., and J. P. Richardson. 1995. NusG is required to overcome a kinetic limitation to Rho function at an intragenic terminator. Proc. Natl. Acad. Sci. USA 92:4738–4742.
28. Burova, E., S. C. Hung, V. Sagitov, B. L. Stitt, and M. E. Gottesman. 1995. Escherichia coli NusG protein stimulates transcription elongation rates in vivo and in vitro. J. Bacteriol. 177:1388–1392.
29. Carafa, Y. D., E. Brody, and C. Thermes. 1990. Prediction of rho-independent Escherichia coli transcription terminators: a statistical analysis of their RNA stem-loop structures. J. Mol. Biol. 216:835–858.
30. Carpousis, A. J., and J. D. Gralla. 1985. Interaction of RNA polymerase with lacUV5 promoter DNA during mRNA initiation and elongation. Footprinting, methylation, and rifampicin-sensitivity changes accompanying transcription initiation. J. Mol. Biol. 183:165–177.
31. Cassani, G., R. R. Burgess, and H. M. Goodman. 1970. Streptolydigin inhibition of RNA polymerase. Cold Spring Harbor Symp. Quant. Biol. 35:59–63.
32. Chamberlin, M. J. 1994. New models for the mechanism of transcription elongation and its regulation. Harvey Lect. 88:1–21.
33. Chamberlin, M. J., and P. Berg. 1962. Deoxyribonucleic acid-directed synthesis of ribonucleic acid by an enzyme from Escherichia coli. Proc. Natl. Acad. Sci. USA 48:81–94.
34. Chan, C. L., and R. Landick. 1989. The Salmonella typhimurium his operon leader region contains an RNA hairpin-dependent transcription pause site. J. Biol. Chem. 264:20796–20804.
35. Chan, C. L., and R. Landick. 1993. Dissection of the his leader pause site by base substitution reveals a multipartite signal that includes a pause RNA hairpin. J. Mol. Biol. 233:25–42.
36. Chan, C. L., and R. Landick. 1994. New perspectives on RNA chain elongation and termination by E. coli RNA polymerase, p. 297–321. In R. C. Conaway and J. W. Conaway (ed.), Transcription: Mechanisms and Regulation. Raven Press, New York.
37. Chattopadhyay, S., J. Garcia-Mena, J. DeVito, K. Wolska, and A. Das. 1995. Bipartite function of a small RNA hairpin in transcription antitermination in bacteriophage λ. Proc. Natl. Acad. Sci. USA 92:4061–4065.
38. Chen, C.-Y. A., G. R. Galluppi, and J. P. Richardson. 1986. Transcription termination at λ tR1 is mediated by interaction of Rho with specific single-stranded domains near the 3' end of cro mRNA. Cell 46:1023–1028.
39. Chen, C.-Y. A., and J. P. Richardson. 1987. Sequence elements essential for ρ-dependent transcription termination at λtR1. J. Biol. Chem. 262:11292–11299.
40. Chenchick, A., R. Beabealashvilli, and A. Mirzabekov. 1981. Topography of interaction of Escherichia coli RNA polymerase subunits with lac UV5 promoter. FEBS Lett. 128:46–50.
41. Cheng, S.-W., E. C. Lynch, K. R. Leason, D. L. Court, B. A. Shapiro, and D. I. Friedman. 1991. Functional importance of sequence in the stem-loop of a transcription terminator. Science 254:1205–1207.
42. Clerget, M., D. J. Jin, and R. A. Weisberg. 1995. A zinc binding region in the β' subunit of RNA polymerase is involved in antitermination of early transcription of phage HK022. J. Mol. Biol. 248:768–780.
43. Craven, M. G., and D. I. Friedman. 1991. Analysis of the Escherichia coli nusA10(Cs) allele: relating nucleotide changes to phenotypes. J. Bacteriol. 173:1485–1491.
44. Cullen, B. R. 1992. Mechanism of action of regulatory proteins encoded by complex retroviruses. Microbiol. Rev. 56:375–394.
45. Darlix, J. L., and F. Fromageot. 1972. Discontinuous in vitro transcription of DNA. Biochimie 54:47–54.
46. Darst, S. A., E. W. Kubalek, and R. D. Kornberg. 1989. Three-dimensional structure of Escherichia coli RNA polymerase holoenzyme determined by electron crystallography. Nature (London) 340:730–732.
47. Das, A. 1993. Control of transcription termination by RNA-binding proteins. Annu. Rev. Biochem. 62:893–930.
48. Das, A., B. Ghosh, S. Barik, and K. Wolska. 1985. Evidence that ribosomal protein S10 itself is a cellular component necessary for transcription antitermination by phage λ N protein. Proc. Natl. Acad. Sci. USA 82:4070–4074.
49. Das, A., M. E. Gottesman, J. Wardwell, P. Trisler, and S. Gottesman. 1983. A mutation in the Escherichia coli ρ gene that inhibits the N protein activity of phage λ. Proc. Natl. Acad. Sci. USA 80:5530–5534.
50. Das, A., and K. Wolska. 1984. Transcription antitermination in vitro by lambda N gene product: requirement for a phage nut site and the product of host nusA, nusB and nusE genes. Cell 38:165–173.
51. Davis, R. W., and R. W. Hyman. 1970. Physical locations of the in vitro RNA initiation site and termination site of T7M DNA. Cold Spring Harbor Symp. Quant. Biol. 35:269–281.
52. de Crombrugghe, B., M. Mudryj, R. Dilauro, and M. Gottesman. 1979. Specificity of the bacteriophage lambda N gene product (pN): nut sequences are necessary and sufficient for antitermination by pN. Cell 18:1145–1151.
53. DeVito, J., and A. Das. 1994. Control of transcription processivity in phage λ: Nus factors strengthen the termination-resistant state of RNA polymerase induced by N antiterminator. Proc. Natl. Acad. Sci. USA 91:8660–8664.
54. Dissinger, S., and M. M. Hanna. 1990. Active site labeling of Escherichia coli transcription elongation complexes with 5-((4-azidophenacyl)thio)uridine 5'-triphosphate. J. Biol. Chem. 265:7662–7668.
55. Dissinger, S., and M. M. Hanna. 1991. RNA-protein interactions in a Nus A-containing Escherichia coli transcription complex paused at an RNA hairpin. J. Mol. Biol. 219:11–25.
56. Doelling, J. H., and N. C. Franklin. 1990. Effects of all single base substitutions in the loop of boxB on antitermination by bacteriophage λ’s N protein. Nucleic Acids Res. 17:5565–5577.
57. Dolan, J. W., N. F. Marshall, and J. P. Richardson. 1990. Transcription termination factor rho has three distinct structural domains. J. Biol. Chem. 265:5747–5754.
58. Dombroski, A. J., C. A. Brennan, P. Spear, and T. Platt. 1988. Site-directed alterations in the ATP-binding domain of rho protein affect its activities as a termination factor. J. Biol. Chem. 263:18802–18809.
59. Dombroski, A. J., J. R. LaDine, R. L. Cross, and T. Platt. 1988. The ATP binding site on rho protein: affinity labeling of Lys181 by pyridoxal 5'-diphospho-5'-adenosine. J. Biol. Chem. 263:18810–18815.
60. Dombroski, A. J., and T. Platt. 1988. Structure of rho factor: an RNA-binding domain and a separate region with strong similarity to proven ATP-binding domains. Proc. Natl. Acad. Sci. USA 85:2538–2542.
61. Dombroski, A. J., and T. Platt. 1990. Mutations in the ATP-binding domain of Escherichia coli rho factor affect transcription termination in vivo. J. Bacteriol. 172:2477–2484.
62. Engel, D., and J. P. Richardson. 1984. Conformational alterations of transcription termination protein rho induced by ATP and by RNA. Nucleic Acids Res. 12:7389–7400.
63. Erie, D. A., O. Hajiseyedjavadi, M. C. Young, and P. H. von Hippel. 1993. Multiple RNA polymerase conformations and GreA: control of the fidelity of transcription. Science 262:867–873.
64. Farnham, P. J., J. Greenblatt, and T. Platt. 1982. Effects of NusA protein on transcription termination in the tryptophan operon of Escherichia coli. Cell 29:945–951.
65. Faus, I., and J. P. Richardson. 1989. Thermodynamic and enzymological characterization of the interaction between transcription termination factor rho and lambda cro mRNA. Biochemistry 28:3510–3517.
66. Faus, I., and J. P. Richardson. 1990. Structural and functional properties of the segments of λcro mRNA that interact with transcription termination factor rho. J. Mol. Biol. 212:53–66.
67. Finger, L. R., and J. P. Richardson. 1981. Procedure for purification of Escherichia coli ribonucleic acid synthesis protein rho. Biochemistry 20:1640–1645.
68. Finger, L. R., and J. P. Richardson. 1982. Stabilization of the hexamer form of Escherichia coli rho under ATP hydrolysis conditions. J. Mol. Biol. 156:203–219.
69. Fisher, R. F., and C. Yanofsky. 1983. Mutations of the β subunit of RNA poly-merase alter both transcription pausing and transcription termination in the trp operon leader region in vitro. J. Biol. Chem. 258:8146–8150.
70. Fisher, R. F., and C. Yanofsky. 1983. A complementary DNA oligomer releases a transcription pause complex. J. Biol. Chem. 258:9208–9212.
71. Forbes, D., and I. Herskowitz. 1982. Polarity suppression by the Q gene product of phage lambda. J. Mol. Biol. 160:549–569.
72. Franklin, N. C. 1974. Altered reading of genetic signals fused to the N operon of bacteriophage lambda: genetic evidence for the modification of polymerase by the protein product of the N gene. J. Mol. Biol. 89:33–48.
73. Franklin, N. C. 1985. Conservation of genome form but not sequence in the transcription antitermination determinants of bacteriophage λ, φ21 and P22. J. Mol. Biol. 181:75–84.
74. Friedman, D. I., and L. S. Baron. 1974. Genetic characterization of a bacterial locus involved in the activity of the N function of phage lambda. Virology 58:141–148.
75. Friedman, D. I., M. Baumann, and L. S. Baron. 1976. Cooperative effects of bacterial mutations affecting lambda N gene expression. I. Isolation and characterization of a nusB mutant. Virology 73:119–127.
76. Friedman, D. I., and M. Gottesman. 1983. Lytic mode of lambda development, p. 21–51. In R. W. Hendrix, J. W. Roberts, F. W. Stahl, and R. A. Weisberg (ed.), Lambda II. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
77. Friedman, D. I., and E. R. Olson. 1983. Evidence that a nucleotide sequence, "boxA," is involved in the action of the NusA protein. Cell 34:143–149.
78. Friedman, D. I., E. R. Olson, L. L. Johnson, D. Alessi, and M. G. Craven. 1990. Transcription-dependent competition for a host factor: the function and optimal sequence of the phage λboxA transcription antitermination signal. Genes Dev. 4:2210–2222.
79. Friedman, D. I., A. T. Sauer, M. R. Baumann, L. S. Baron, and S. L. Adhya. 1981. Evidence that ribosomal protein S10 participates in the control of transcription termination. Proc. Natl. Acad. Sci. USA 78:1115–1118.
80. Friedman, D. I., G. S. Wilgus, and R. J. Mural. 1973. Gene N regulator function of phage lambda imm21: evidence that a site of N action differs from a site of N recognition. J. Mol. Biol. 81:505–516.
81. Galluppi, G. R., and J. P. Richardson. 1980. ATP-induced changes in the binding of RNA synthesis termination protein rho to RNA. J. Mol. Biol. 138:513–539.
82. Gamper, H. B., and J. E. Hearst. 1982. A topological model for transcription based on unwinding angle analysis of E. coli RNA polymerase binary, initiation and ternary complexes. Cell 29:81–90.
83. Geiduschek, E. P., T. Nakamoto, and S. B. Weiss. 1961. The enzymatic synthesis of RNA: complementary interaction with DNA. Proc. Natl. Acad. Sci. USA 47:1405–1415.
84. Geiselmann, J., S. E. Seifried, T. D. Yager, C. Liang, and P. H. von Hippel. 1992. Physical properties of the Escherichia coli transcription termination factor rho. 2. Quaternary structure of the rho hexamer. Biochemistry 31:121–132.
85. Geiselmann, J., and P. H. von Hippel. 1992. Functional interactions of ligand cofactors with Escherichia coli transcription termination factor rho. I. Binding of ATP. Protein Sci. 1:850–860.
86. Geiselmann, J., Y. Wang, S. E. Seifried, and P. H. von Hippel. 1993. A physical model for the translocation and helicase activities of Escherichia coli transcription termination protein Rho. Proc. Natl. Acad. Sci. USA 90:7754–7758.
87. Geiselmann, J., T. D. Yager, S. C. Gill, P. Calmettes, and P. H. von Hippel. 1992. Physical properties of the Escherichia coli transcription termination factor rho. 1. Association states and geometry of the rho hexamer. Biochemistry 31:111–121.
88. Geiselmann, J., T. D. Yager, and P. H. von Hippel. 1992. Functional interactions of ligand cofactors with Escherichia coli transcription termination factor rho. II. Binding of RNA. Protein Sci. 1:861–873.
89. Georgopoulos, C. P. 1971. Bacterial mutants in which the gene N function of bacteriophage lambda is blocked have an altered RNA polymerase. Proc. Natl. Acad. Sci. USA 68:2977–2981.
90. Georgopoulos, C. P., J. Swindle, F. Keppel, M. Ballivet, H. Bisig, and H. Eisen. 1980. Studies on the E. coli groNB(nusB) gene which affects bacteriophage λN gene function. Mol. Gen. Genet. 179:55–61.
91. Ghosh, B., and A. Das. 1984. NusB: a protein factor necessary for transcription antitermination in vitro by phage λN gene product. Proc. Natl. Acad. Sci. USA 81:6305–6309.
92. Ghysen, A., and M. Pironio. 1972. Relationship between the N function of bacteriophage lambda and host RNA polymerase. J. Mol. Biol. 65:259–272.
93. Gill, S. C., S. E. Weitzel, and P. H. von Hippel. 1991. Escherichia coli σ70 and NusA proteins I. Binding interactions with core RNA polymerase in solution and within the transcription complex. J. Mol. Biol. 220:307–324.
94. Gish, K., and C. Yanofsky. 1993. Inhibition of expression of the tryptophanase operon in Escherichia coli by extrachromosomal copies of the tna leader region. J. Bacteriol. 175:3380–3387.
95. Goda, Y., and J. Greenblatt. 1985. Efficient modification of E. coli RNA poly-merase in vitro by the N gene transcription antitermination protein of bacteriophage lambda. Nucleic Acids Res. 13:2569–2582.
96. Gogol, E. P., S. E. Seifried, and P. H. von Hippel. 1991. Structure and assembly of the Escherichia coli transcription termination factor rho and its interactions with RNA. I. Cryoelectron microscopic studies. J. Mol. Biol. 221:1127–1138.
97. Gottesman, S., M. Gottesman, J. E. Shaw, and M. L. Pearson. 1991. Protein degradation in E. coli: the lon mutation and bacteriophage lambda N and cII protein stability. Cell 24:225–233.
98. Grayhack, E. J., and J. W. Roberts. 1982. The phage λ gene Q product: activity of a transcription antiterminator in vitro. Cell 30:637–648.
99. Grayhack, E. J., X. Yang, L. F. Lau, and J. W. Roberts. 1985. Phage lambda gene Q antiterminator recognizes RNA polymerase near the promoter and accelerates it through a pause site. Cell 42:259–269.
100. Greenblatt, J. 1972. Positive control of endolysin synthesis in vitro by the gene N protein of phage lambda. Proc. Natl. Acad. Sci. USA 69:3606–3610.
101. Greenblatt, J. 1973. Regulation of the expression of the N gene of bacteriophage lambda. Proc. Natl. Acad. Sci. USA 70:421–424.
102. Greenblatt, J. 1981. Regulation of transcription termination by the N gene protein of bacteriophage lambda. Cell 24:8–9.
103. Greenblatt, J. 1984. Regulation of transcription termination in E. coli. Can. J. Biochem. Cell. Biol. 62:79–88.
104. Greenblatt, J., and J. Li. 1981. The nusA gene product of Escherichia coli; its identification and a demonstration that it interacts with the gene N transcription antitermination protein of bacteriophage lambda. J. Mol. Biol. 147:11–23.
105. Greenblatt, J., and J. Li. 1981. Interaction of the sigma factor and the nusA gene protein of E. coli with RNA polymerase in the initiation-termination cycle of transcription. Cell 24:421–428.
106. Greenblatt, J., J. Li, S. Adhya, D. I. Friedman, L. S. Baron, B. Redfield, H.-F. Kung, and H. Weissbach. 1980. L factor that is required for β-galactosidase synthesis is the nusA gene product involved in transcription termination. Proc. Natl. Acad. Sci. USA 77:1991–1994.
107. Greenblatt, J., P. Malnoe, and J. Li. 1980. Purification of the gene N transcription antitermination protein of bacteriophage λ. J. Biol. Chem. 255:1465–1470.
108. Greenblatt, J., M. McLimont, and S. Hanly. 1981. Termination of transcription by the nusA gene protein of Escherichia coli. Nature (London) 292:215–220.
109. Greenblatt, J., J. R. Nodwell, and S. W. Mason. 1993. Transcriptional antitermination. Nature (London) 364:401–406.
110. Hanna, M. M., and C. F. Meares. 1983. Topography of transcription: path of the leading end of nascent RNA through the Escherichia coli transcription complex. Proc. Natl. Acad. Sci. USA 80:4238–4242.
111. Hansen, U. M., and W. R. McClure. 1980. Role of the σ subunit of Escherichia coli RNA polymerase in initiation. II. Release of σ from ternary complexes. J. Biol. Chem. 255:9564–9570.
112. Hart, C. M., and J. W. Roberts. 1991. Rho-dependent transcription termination. Characterization of the requirement for cytidine in the nascent transcript. J. Biol. Chem. 266:24140–24148.
113. Hart, C. M., and J. W. Roberts. 1994. Deletion analysis of the lambda tR1 termination region: effect of sequences near the transcript release sites, and the minimum length of Rho-dependent transcripts. J. Mol. Biol. 237:255–265.
114. Heisler, L. M., H. Suzuki, R. Landick, and C. A. Gross. 1993. Four contiguous amino acids define the target for streptolydigin resistance in the β subunit of Escherichia coli RNA polymerase. J. Biol. Chem. 268:25369–25375.
115. Hinkle, D. C., and M. J. Chamberlin. 1972. Studies of the binding of Escherichia coli RNA polymerase to DNA. I. The role of sigma subunit in site selection. J. Mol. Biol. 70:157–185.
116. Holben, W. E., and E. A. Morgan. 1984. Antitermination of transcription from an Escherichia coli ribosomal RNA promoter. Proc. Natl. Acad. Sci. USA 81:6789–6793.
117. Horowitz, D. S., and A. R. Kraimer. 1994. Mechanisms for selecting 5' splice sites in mammalian pre-mRNA splicing. Trends Genet. 10:100–108.
118. Horwitz, R. J., J. Li, and J. Greenblatt. 1987. An elongation control particle containing the N gene transcription antitermination protein of bacteriophage lambda. Cell 51:631–641.
119. Houman, F., M. R. Diaz-Torres, and A. Wright. 1990. Transcriptional antitermination in the bgl operon of E. coli is modulated by a specific RNA binding protein. Cell 62:1153–1163.
120. Howard, B., and B. deCrombrugghe. 1976. ATPase activity required for termination of transcription by the Escherichia coli protein factor rho. J. Biol. Chem. 251:2520–2524.
121. Hung, S. C., and M. E. Gottesman. 1995. Phage HK022 Nun protein arrests transcription on phage λ DNA in vitro and competes with the phage λ N antitermination protein. J. Mol. Biol. 247:428–442.
122. Isaksen, M. L., S. T. Rishovd, R. Calendar, and B. H. Lindqvist. 1992. The polarity suppression factor of bacteriophage P4 is also a decoration protein of the P4 capsid. Virology 188:831–839.
123. Ishii, S., M. Ihara, T. Maekawa, H. Nakamura, H. Uchida, and F. Imamoto. 1984. The nucleotide sequence of the cloned nusA gene and its flanking region of Escherichia coli. Nucleic Acids Res. 12:3333–3342.
124. Ishii, S., J. Salstrom, Y. Sugino, W. Szybalski, and F. Imamoto. 1980. A biochemical assay for the transcription-antitermination function of the coliphage λN gene product. Gene 10:17–25.
125. Jin, D. J., R. R. Burgess, J. P. Richardson, and C. A. Gross. 1992. Termination efficiency at rho-dependent terminators depends on kinetic coupling between RNA polymerase and rho. Proc. Natl. Acad. Sci. USA 89:1453–1457.
126. Jin, D. J., and C. A. Gross. 1988. Mapping and sequencing of mutations in the Escherichia coli rpoB gene that lead to rifampin resistance. J. Mol. Biol. 202:45–58.
127. Jin, D. J., and C. A. Gross. 1991. RpoB8, a rifampicin-resistant termination-proficient RNA polymerase, has an increased Km for purine nucleotides during transcription elongation. J. Biol. Chem. 266:14478–14485.
128. Jin, D. J., W. A. Walter, and C. A. Gross. 1988. Characterization of termination phenotypes of rifampicin-resistant mutants. J. Mol. Biol. 202:245–253.
129. Johnston, D. E., and W. R. McClure. 1976. Abortive initiation of in vitro RNA synthesis on bacteriophage λ DNA, p. 413–428. In R. Losick and M. Chamberlin (ed.), RNA Polymerase. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.
130. Kainz, M., and J. Roberts. 1992. Structure of transcription elongation complexes in vivo. Science 255:838–841.
131. Kameyama, L., L. Fernandez, D. L. Court, and G. Guarneros. 1991. RNase III activation of bacteriophage λ N synthesis. Mol. Microbiol. 5:2953–2963.
132. Kashlev, M., J. Lee, K. Zalenskaya, V. Nikiforov, and A. Goldfarb. 1990. Blocking of the initiation-to-elongation transition by a transdominant RNA polymerase mutation. Science 248:1006–1009.
133. Kassavetis, G. A., and M. J. Chamberlin. 1981. Pausing and termination of transcription within the early region of bacteriophage T7 DNA in vitro. J. Biol. Chem. 256:2777–2786.
134. Kenan, D. J., C. C. Query, and J. D. Keene. 1991. RNA recognition: towards identifying determinants of specificity. Trends Biochem. Sci. 16:214–220.
135. Keppel, F., C. Georgopoulos, and H. Eisen. 1974. Host interference with expression of the lambda N gene product. Biochimie 56:1503–1509.
136. Kingston, R. E., and M. J. Chamberlin. 1981. Pausing and attenuation of in vitro transcription in the rrnB operon of E. coli. Cell 27:523–531.
137. Konrad, M. 1968. Dependence of "early" λ bacteriophage RNA synthesis on bacteriophage-directed protein synthesis. Proc. Natl. Acad. Sci. USA 59:171–178.
138. Krummel, B., and M. J. Chamberlin. 1992. Structural analysis of ternary complexes of Escherichia coli RNA polymerase. Individual complexes halted along different transcription units have distinct and unexpected biochemical properties. J. Mol. Biol. 225:221–237.
139. Krummel, B., and M. J. Chamberlin. 1992. Structural analysis of ternary complexes of Escherichia coli RNA polymerase. Deoxyribonuclease I footprinting of defined complexes. J. Mol. Biol. 225:239–250.
140. Kumar, S. A., and J. S. Krakow. 1975. Studies on the product binding site of the Azotobacter vinelandii ribonucleic acid polymerase. J. Biol. Chem. 250:2878–2884.
141. Kung, H.-F., C. Spears, and H. Weissbach. 1975. Purification and properties of a soluble factor required for the deoxyribonucleic acid-directed in vitro synthesis of β-galactosidase. J. Biol. Chem. 250:1556–1562.
142. Landick, R., J. Carey, and C. Yanofsky. 1985. Translation activates the paused transcription complex and restores transcription of the trp operon leader region. Proc. Natl. Acad. Sci. USA 82:4663–4667.
143. Landick, R., J. Stewart, and D. N. Lee. 1990. Amino acid changes in conserved regions of the β-subunit of Escherichia coli RNA polymerase alter transcription pausing and termination. Genes Dev. 4:1623–1636.
144. Landick, R., and C. Yanofsky. 1984. Stability of an RNA secondary structure affects in vitro transcription pausing in the trp operon leader region. J. Biol. Chem. 259:11550–11555.
145. Landick, R., and C. Yanofsky. 1987. Isolation and structural analysis of the Escherichia coli trp leader paused transcription complex. J. Mol. Biol. 196:363–377.
146. Landick, R., and C. Yanofsky. 1987. Transcription attenuation, p. 1276–1301. In F. C. Neidhardt, J. L. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, vol. 2. American Society for Microbiology, Washington, D.C.
147. Lau, L. F., J. W. Roberts, and R. Wu. 1982. Transcription terminates at λtR1 in three clusters. Proc. Natl. Acad. Sci. USA 79:6171–6175.
148. Lau, L. F., J. W. Roberts, and R. Wu. 1983. RNA polymerase pausing and transcript release at the λ tR1 terminator in vitro. J. Biol. Chem. 258:9391–9397.
149. Lau, L. F., J. W. Roberts, R. Wu, F. Georges, and S. A. Narang. 1984. A potential stem-loop structure and the sequence CAAUCAA in the transcript are insufficient to signal ρ-dependent transcription termination at λ tR1. Nucleic Acids Res. 12:1287–1299.
150. Lazinski, D., E. Grzadzielska, and A. Das. 1989. Sequence-specific recognition of RNA hairpins by bacteriophage antiterminators requires a conserved arginine-rich motif. Cell 59:207–218.
151. Lee, D. N., and R. Landick. 1992. Structure of RNA and DNA chains in paused transcription complexes containing Escherichia coli RNA polymerase. J. Mol. Biol. 228:759–777.
152. Lee, D. N., L. Phung, J. Stewart, and R. Landick. 1990. Transcription pausing by Escherichia coli RNA polymerase is modulated by downstream DNA sequences. J. Biol. Chem. 265:15145–15153.
153. Levin, J. R., and M. J. Chamberlin. 1987. Mapping and characterization of transcriptional pause sites in the early genetic region of bacteriophage T7. J. Mol. Biol. 196:61–84.
154. Levin, J. R., B. Krummel, and M. J. Chamberlin. 1987. Isolation and properties of transcribing ternary complexes of Escherichia coli RNA polymerase positioned at a single template base. J. Mol. Biol. 196:85–100.
155. Li, J., R. Horwitz, S. McCracken, and J. Greenblatt. 1992. NusG, a new Escherichia coli elongation factor involved in transcription antitermination by the N protein of phage lambda. J. Biol. Chem. 267:6012–6019.
156. Li, J., S. W. Mason, and J. Greenblatt. 1993. Elongation factor NusG interacts with termination factor ρ to regulate termination and antitermination of transcription. Genes Dev. 7:161–172.
157. Li, S. C., C. L. Squires, and C. Squires. 1984. Antitermination of E. coli rRNA transcription is caused by a control region segment containing lambda nut-like sequences. Cell 38:851–860.
158. Linderoth, N. A., and R. L. Calendar. 1991. The Psu protein of bacteriophage P4 is an antitermination factor for rho-dependent transcription termination. J. Bacteriol. 173:6722–6731.
159. Linn, T., and J. Greenblatt. 1992. The NusA and NusG proteins of Escherichia coli increase the in vitro readthrough frequency of a transcriptional attenuator preceding the gene for the β subunit of RNA polymerase. J. Biol. Chem. 267:1449–1454.
160. Liu, B., and B. M. Alberts. 1995. Head-on collision between a DNA replication apparatus and RNA polymerase transcription complex. Science 267:1131–1137.
161. Liu, B., M. L. Wong, and B. M. Alberts. 1994. A transcribing RNA polymerase molecule survives DNA replication without aborting its growing RNA chain. Proc. Natl. Acad. Sci. USA 91:10660–10664.
162. Liu, K., and M. M. Hanna. 1995. NusA contacts RNA in Escherichia coli transcription complexes. J. Mol. Biol. 247:547–558.
163. Lowery, C., and J. P. Richardson. 1977. Characterization of the nucleoside triphosphate phosphohydrolase (ATPase) activity of RNA synthesis termination factor ρ.I. Enzymatic properties and effects of inhibitors. J. Biol. Chem. 252:1375–1380.
164. Lowery, C., and J. P. Richardson. 1977. Characterization of the nucleoside triphosphate phosphohydrolase (ATPase) activity of RNA synthesis termination factor ρ. II. Influence of synthetic RNA homopolymers and random copolymers on the reaction. J. Biol. Chem. 252:1381–1385.
165. Lowery-Goldhammer, C., and J. P. Richardson. 1974. An RNA-dependent nucleoside triphosphate phosphohydrolase (ATPase) associated with rho termination factor. Proc. Natl. Acad. Sci. USA 71:2003–2007.
166. Lozeron, H. A., P. J. Anevski, and D. Apirion. 1977. Antitermination and absence of processing of the leftward transcript of coliphage lambda in the RNAse III-deficient host. J. Mol. Biol. 109:359–365.
167. Luzzati, D. 1970. Regulation of λ exonuclease synthesis: role of the N gene product and λ repressor. J. Mol. Biol. 49:515–519.
168. Lynn, S. P., L. M. Kasper, and J. F. Gardner. 1988. Contributions of RNA secondary structure and length of the thymidine tract to transcription termination at the thr operon attenuator. J. Biol. Chem. 263:472–479.
169. Mahadevan, S., and A. Wright. 1987. A bacterial gene involved in transcription antitermination: regulation at a rho-independent terminator in the bgl operon of E. coli. Cell 50:485–494.
170. Martin, F. H., and I. J. Tinoco. 1980. DNA-RNA hybrid duplexes containing oligo(dA:rU) sequences are exceptionally unstable and may facilitate termination of transcription. Nucleic Acids Res. 8:2295–2299.
171. Mason, S., and J. Greenblatt. 1991. Assembly of transcription elongation complexes containing the N protein of phage λ and the Escherichia coli elongation factors NusA, NusB, NusG and S10. Genes Dev. 5:1504–1512.
172. Mason, S., J. Li, and J. Greenblatt. 1992. A direct interaction between two Escherichia coli transcription antitermination factors, NusB and ribosomal protein S10. J. Mol. Biol. 223:55–66.
173. Mason, S. W., J. Li, and J. Greenblatt. 1992. Host factor requirements for processive antitermination of transcription and suppression of pausing by the N protein of bacteriphage λ. J. Biol. Chem. 267:19418–19426.
174. McDowell, J. C., J. W. Roberts, D. J. Jin, and C. Gross. 1994. Determination of intrinsic transcription termination efficiency by RNA polymerase elongation rate. Science 266:822–825.
175. McSwiggen, J. A., D. G. Bear, and P. H. von Hippel. 1988. Interactions of Escherichia coli termination factor rho with RNA. I. Binding stoichiometries and free energies. J. Mol. Biol. 199:609–622.
176. Metzger, W., P. Schickor, and H. Heumann. 1989. A cinematographic view of Escherichia coli RNA polymerase translocation. EMBO J. 8:2745–2754.
177. Michalke, H., and H. Bremer. 1969. RNA synthesis in Escherichia coli after irradiation with ultraviolet light. J. Mol. Biol. 41:1–23.
178. Milman, G., R. Langridge, and M. J. Chamberlin. 1967. The structure of a DNA-RNA hybrid. Proc. Natl. Acad. Sci. USA 57:1804–1810.
179. Modrak, D., and J. P. Richardson. 1994. The RNA-binding domain of transcription termination factor Rho: isolation, characterization and determination of sequence limits. Biochemistry 33:8292–8299.
180. Morgan, E. A. 1980. Insertions of Tn10 into an E. coli ribosomal RNA operon are incompletely polar. Cell 21:257–265.
181. Morgan, W. D., D. G. Bear, B. L. Litchman, and P. H. von Hippel. 1985. RNA sequence and secondary structure requirements for rho-dependent transcription termination. Nucleic Acids Res. 13:3739–3754.
182. Morgan, W. D., D. G. Bear, and P. H. von Hippel. 1983. Rho-dependent termination of transcription. I. Identification and characterization of termination sites for transcription from the bacteriophage λ PR promoter. J. Biol. Chem. 258:9553–9564.
183. Morgan, W. D., D. G. Bear, and P. H. von Hippel. 1983. Rho-dependent termination of transcription. II. Kinetics of mRNA elongation during transcription from the bacteriophage λ PR promoter. J. Biol. Chem. 258:9565–9574.
184. Morgan, W. D., D. G. Bear, and P. H. von Hippel. 1984. Specificity of release by Escherichia coli transcription termination factor Rho of nascent mRNA transcripts initiated at the λ PR promoter. J. Biol. Chem. 259:8664–8671.
185. Mosteller, R. D., and C. Yanofsky. 1970. Transcription of the tryptophan operon in Escherichia coli: rifampicin as an inhibitor of initiation. J. Mol. Biol. 48:525–531.
186. Neff, N. F., and M. J. Chamberlin. 1980. Termination of transcription by Escherichia coli RNA polymerase in vitro. Effect of altered reaction conditions and mutations in the enzyme protein on termination with T7 and T3 DNAs. Biochemistry 19:3005–3015.
187. Nehrke, K. W., and T. Platt. 1994. A quaternary transcription termination complex. Reciprocal stabilization by Rho factor and NusG protein. J. Mol. Biol. 243:830–839.
188. Nodwell, J. R., and J. Greenblatt. 1991. The nut site of bacteriophage λ is made of RNA and is bound by transcription antitermination factors on the surface of RNA polymerase. Genes Dev. 5:2141–2151.
189. Nodwell, J. R., and J. Greenblatt. 1993. Recognition of boxA antiterminator RNA by the E. coli antitermination factors NusB and ribosomal protein S10. Cell 72:261–268.
190. Nudler, E., A. Goldfarb, and M. Kashlev. 1994. Discontinuous mechanism of transcription elongation. Science 265:793–796.
191. Nudler, E., M. Kashlev, V. Nikifovov, and A. Goldfarb. 1995. Coupling between transcription termination and RNA polymerase inchworming. Cell 81:351–358.
192. Oberto, J., M. Clerget, M. Ditto, K. Cam, and R. A. Weisberg. 1993. Antitermination of early transcription in phage HK022. Absence of a phage-encoded antitermination factor. J. Mol. Biol. 229:368–381.
193. Oberto, J., R. A. Weisberg, and M. E. Gottesman. 1989. Structure and function of the nun gene and the immunity region of the lambdoid phage HK022. J. Mol. Biol. 207:675–693.
194. Oda, T., and M. Takanami. 1972. Observations on the structure of the termination factor rho and its attachment to DNA. J. Mol. Biol. 71:799–802.
195. Olson, E. R., E. L. Flamm, and D. I. Friedman. 1982. Analysis of nutR: a region of phage lambda required for antitermination of transcription. Cell 31:61–70.
196. Olson, E. R., C.-S. C. Tomich, and D. I. Friedman. 1984. The nusA recognition site: alteration in its sequence or position relative to upstream translation interferes with the action of the N antitermination function of phage λ. J. Mol. Biol. 180:1053–1063.
197. Opperman, T., and J. P. Richardson. 1994. A phylogenetic analysis of sequences from diverse bacteria with homology to the Escherichia coli rho gene. J. Bacteriol. 176:5033–5043.
198. Orlova, M., J. Newlands, A. Das, A. Goldfarb, and S. Borukhov. 1995. Intrinsic transcription cleavage activity of RNA polymerase. Proc. Natl. Acad. Sci. USA 92:4596–4600.
199. Patterson, T. A., Z. Zhang, T. Baker, L. L. Johnson, D. I. Friedman, and D. Court. 1994. Bacteriophage lambda N-dependent transcription antitermination. Competition for an RNA site may regulate antitermination. J. Mol. Biol. 236:217–228.
200. Peltz, S. W., A. L. Brown, N. Hasan, A. J. Podhajska, and W. Szybalski. 1985. Thermosensitivity of a DNA recognition site: activity of a truncated nutL antiterminator of coliphage lambda. Science 228:91–93.
201. Pinkham, J. L., and T. Platt. 1983. The nucleotide sequence of the rho gene of E. coli K-12. Nucleic Acids Res. 11:3531–3545.
202. Platt, T. 1994. Rho and RNA: models for recognition and response. Mol. Microbiol. 11:983–990.
203. Platt, T., and J. P. Richardson. 1992. Escherichia coli Rho factor: protein and enzyme of transcription termination, p. 365–388. In S. L. McKnight and K. R. Yamamoto (ed.), Transcriptional Regulation. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
204. Query, C. C., R. C. Bentley, and J. D. Keene. 1989. A common RNA recognition motif identified within a defined U1 RNA binding domain of the 70K U1 snRNP protein. Cell 57:89–101.
205. Rajapandi, T., and D. Oliver. 1994. ssaD1, a suppressor of secA51(Ts) that renders growth of Escherichia coli cold sensitive, is an early amber mutation in the transcription factor gene nusB. J. Bacteriol. 176:4444–4447.
206. Rees, W. A., R. W. Keller, J. P. Vesenka, G. Yang, and C. Bustamante. 1993. Evidence of DNA bending in transcription complexes imaged by scanning force microscopy. Science 260:1646–1649.
207. Reynolds, R., and M. J. Chamberlin. 1992. Parameters affecting transcription termination by Escherichia coli RNA polymerase. II. Construction and analysis of hybrid terminators. J. Mol. Biol. 224:53–63.
208. Rice, G. A., C. M. Kane, and M. J. Chamberlin. 1991. Footprinting analysis of mamalian RNA polymerase II along its transcript: an alternative view of transcription elongation. Proc. Natl. Acad. Sci. USA 88:4245–4249.
209. Richardson, J. P. 1970. Rho factor function in T4 RNA transcription. Cold Spring Harbor Symp. Quant. Biol. 35:127–133.
210. Richardson, J. P. 1982. Activation of rho protein ATPase requires simultaneous interaction at two kinds of nucleic acid-binding sites. J. Biol. Chem. 257:5760–5766.
211. Richardson, J. P. 1990. Rho-dependent transcription termination. Biochim. Biophys. Acta 1048:127–138.
212. Richardson, J. P. 1991. Preventing the synthesis of unused transcripts by rho factor. Cell 64:1047–1049.
213. Richardson, J. P. 1993. Transcription termination. Crit. Rev. Biochem. Mol. Biol. 28:1–30.
214. Richardson, J. P., and J. L. Carey. 1982. rho factors from polarity suppressor mutants with defects in their RNA interactions. J. Biol. Chem. 257:5767–5771.
215. Richardson, J. P., and R. Conaway. 1980. Ribonucleic acid release activity of transcription termination protein ρ is dependent on the hydrolysis of nucleoside triphosphates. Biochemistry 19:4293–4299.
216. Richardson, J. P., C. Grimley, and C. Lowery. 1975. Transcription termination factor rho activity is altered in Escherichia coli with suA gene mutations. Proc. Natl. Acad. Sci. USA 72:1725–1728.
217. Ring, B. Z., and J. W. Roberts. 1994. Function of a nontranscribed DNA strand site in transcription elongation. Cell 78:317–324.
218. Robert, J., S. B. Sloan, R. A. Weisberg, M. E. Gottesman, R. Robledo, and D. Harbrecht. 1987. The remarkable specificity of a new transcription termination factor suggests that the mechanisms of termination and antitermination are similar. Cell 51:483–492.
219. Roberto, R., B. L. Atkinson, and M. E. Gottesman. 1991. Escherichia coli mutations that block transcription termination by phage HK022 Nun protein. J. Mol. Biol. 220:613–619.
220. Roberts, J. W. 1969. Termination factor for RNA synthesis. Nature (London) 224:1168–1174.
221. Roberts, J. W. 1988. Phage lambda and the regulation of transcription termination. Cell 52:5–6.
222. Robledo, R., M. E. Gottesman, and R. A. Weisberg. 1990. λnutR mutations convert HK022 Nun protein from a transcription termination factor to a suppressor of termination. J. Mol. Biol. 212:635–643.
223. Rosenberg, M., and D. Court. 1979. Regulatory sequences involved in the promotion and termination of RNA transcription. Annu. Rev. Genet. 13:319–353.
224. Rosenberg, M., D. Court, H. Shimatake, C. Brady, and D. L. Wulff. 1978. The relationship between function and DNA sequence in an intercistronic regulatory region in phage λ. Nature (London) 272:414–423.
225. Ross, W., K. K. Gosink, J. Salomon, K. Igarashi, C. Zou, A. Ishihama, K. Severinov, and R. L. Gourse. 1993. A third recognition element in bacterial promoters: DNA binding by the α subunit of RNA polymerase. Science 262:1407–1413.
226. Ruteshouser, E. C., and J. P. Richardson. 1989. Identification and characterization of transcription termination sites in the Escherichia coli lacZ gene. J. Mol. Biol. 208:23–43.
227. Ryan, T., and M. J. Chamberlin. 1983. Transcription analyses with heteroduplex trp attenuator templates indicate that the transcript stem and loop structure serves as the termination signal. J. Biol. Chem. 258:4690–4693.
228. Salstrom, J. S., and W. Szybalski. 1978. Coliphage λnutL –: a unique class of mutants defective in the site of gene N product utilization for antitermination of leftward transcription. J. Mol. Biol. 124:195–221.
229. Sasse-Dwight, S., and J. D. Gralla. 1989. KMnO4 as a probe for lac promoter DNA melting and mechanism in vivo. J. Biol. Chem. 264:8074–8081.
230. Schauer, A. T., D. L. Carver, B. Bigelow, L. S. Baron, and D. I. Friedman. 1987. λ N antitermination system: functional analsysis of phage interactions with the host NusA protein. J. Mol. Biol. 194:679–690.
231. Schmidt, M. C., and M. J. Chamberlin. 1984. Amplification and isolation of Escherichia coli NusA protein and studies of its effects on in vitro RNA chain elongation. Biochemistry 23:197–203.
232. Schmidt, M. C., and M. J. Chamberlin. 1984. Binding of rho factor to Escherichia coli RNA polymerase mediated by NusA protein. J. Biol. Chem. 259:15000–15002.
233. Schmidt, M. C., and M. J. Chamberlin. 1987. NusA protein of Escherichia coli is an efficient transcription termination factor for certain termination sites. J. Mol. Biol. 195:809–818.
234. Schwartz, M. 1970. On the function of the N cistron in phage λ. Virology 40:23–33.
235. Segawa, T., and F. Imamoto. 1974. Diversity of regulation of genetic transcription. II. Specific relaxation of polarity in read-through transcription of the translocated trp operon in bacteriophage lambda trp. J. Mol. Biol. 87:741–754.
236. Seifried, S. E., K. P. Bjornson, and P. H. von Hippel. 1991. Structure and assembly of the Escherichia coli transcription termination factor rho and its interactions with RNA. II. Physical chemical studies. J. Mol. Biol. 221:1139–1151.
237. Seifried, S. E., J. B. Easton, and P. H. von Hippel. 1992. ATPase activity of transcription-termination factor rho: functional dimer model. Proc. Natl. Acad. Sci. USA 89:10454–10458.
238. Selby, C. P., and A. Sancar. 1991. Gene- and strand-specific repair in vitro: partial purification of a transcription-repair coupling factor. Proc. Natl. Acad. Sci. USA 88:8232–8236.
239. Selby, C. P., and A. Sancar. 1993. Molecular mechanism of transcription-repair coupling. Science 260:53–58.
240. Selby, C. P., and A. Sancar. 1993. Transcription-repair coupling and mutation frequency decline. J. Bacteriol. 175:7509–7514.
241. Sha, Y., L. Lindahl, and J. M. Zengel. 1995. RNA determinants required for L4-mediated attenuation control of the S10 r-protein operon of Escherichia coli. J. Mol. Biol. 245:486–498.
242. Sha, Y., L. Lindahl, and J. M. Zengel. 1995. Role of NusA in L4-mediated attenuation control of the S10 r-protein operon of Escherichia coli. J. Mol. Biol. 245:474–485.
243. Sharrock, R. A., R. L. Gourse, and M. Nomura. 1985. Defective antitermination of rRNA transcription and derepression of rRNA and tRNA synthesis in the nusB5 mutant of Escherichia coli. Proc. Natl. Acad. Sci. USA 82:5275–5279.
244. Shigesada, K., and C.-W. Wu. 1980. Studies of RNA release reaction catalyzed by E. coli transcription termination factor rho using isolated ternary transcription complexes. Nucleic Acids Res. 8:3355–3369.
245. Simpson, R. B. 1979. The molecular topography of RNA polymerase-promoter interaction. Cell 18:277–285.
246. Sippel, A., and G. Hartmann. 1968. Mode of action of rifamycin on the RNA polymerase reaction. Biochim. Biophys. Acta 157:218–219.
247. Sloan, S. B., and R. A. Weisberg. 1993. Use of a gene encoding a suppressor tRNA as a reporter of transcription: analyzing the action of the Nun protein of bacteriophage HK022. Proc. Natl. Acad. Sci. USA 90:9842–9846.
248. Somasekhar, G., and W. Szybalski. 1983. Mapping of the Q utilization site (qut) required for antitermination of late transcription in bacteriophage lambda. Gene 26:291–294.
249. Sousa, R., Y. J. Chung, J. P. Rose, and B.-C. Wang. 1993. Crystal structure of bacteriophage T7 RNA polymerase at 3.3Å resolution. Nature (London) 364:593–599.
250. Squires, C. L., J. Greenblatt, J. Li, C. Condon, and C. L. Squires. 1993. Ribosomal RNA antitermination in vitro: requirement for Nus factors and one or more unidentified cellular components. Proc. Natl. Acad. Sci. USA 90:970–974.
251. Stebbins, C. E., S. Borukhov, M. Orlova, A. Polyakov, A. Goldfarb, and S. A. Darst. 1995. Crystal structure of the GreA transcript cleavage factor from Escherichia coli. Nature (London) 373:636–640.
252. Steege, D. A., K. C. Cone, C. Queen, and M. Rosenberg. 1987. Bacteriophage λN gene leader RNA. RNA processing and translational initiation signals. J. Biol. Chem. 262:17651–17658.
253. Steinmetz, E. J., C. A. Brennan, and T. Platt. 1990. A short intervening structure can block rho factor helicase action at a distance. J. Biol. Chem. 265:18408–18413.
254. Steinmetz, E. J., and T. Platt. 1994. Evidence supporting a tethered tracking model for helicase activity of Escherichia coli Rho factor. Proc. Natl. Acad. Sci. USA 91:1401–1405.
255. Stewart, V., and C. Yanofsky. 1985. Evidence for transcription antitermination control of tryptophanase operon expression in Escherichia coli K-12. J. Bacteriol. 164:731–740.
256. Stillman, B. 1994. Smart machines at the DNA replication fork. Cell 78:725–728.
257. Stitt, B. L. 1988. Escherichia coli transcription termination protein rho has three hydrolytic sites for ATP. J. Biol. Chem. 263:11130–11137.
258. Straney, D. C., and D. M. Crothers. 1985. Intermediates in transcription initiation from the E. coli lac UV5 promoter. Cell 43:449–459.
259. Sullivan, S. L., and M. E. Gottesman. 1992. Requirement for E. coli NusG protein in factor-dependent transcription termination. Cell 68:989–994.
260. Sullivan, S. L., D. F. Ward, and M. E. Gottesman. 1992. Effect of Escherichia coli nusG function on λ N-mediated transcription antitermination. J. Bacteriol. 174:1339–1344.
261. Swindle, J., C. Zylicz, C. Georgopoulos, J. Li, and J. Greenblatt. 1988. Purification and properties of the NusB protein of Escherichia coli. J. Biol. Chem. 263:10229–10235.
262. Telesnitsky, A., and M. J. Chamberlin. 1989. Terminator-distal sequences determine the in vitro efficiency of the early terminators of bacteriophages T3 and T7. Biochemistry 28:5210–5218.
263. Tsurushita, N., K. Shigesada, and M. Imai. 1989. Mutant rho factors with increased transcription termination activities. I. Functional correlations of the primary and secondary polynucleotide binding sites with the efficiency and site-selectivity of rho-dependent termination. J. Mol. Biol. 210:23–37.
264. Vogel, U., and K. F. Jensen. 1994. The RNA chain elongation rate in Escherichia coli depends on the growth rate. J. Bacteriol. 176:2807–2813.
265. von Hippel, P. H., and T. D. Yager. 1991. Transcription elongation and termination are competitive kinetic processes. Proc. Natl. Acad. Sci. USA 88:2307–2311.
266. Wang, D., T. I. Meier, C. L. Chan, G. Feng, D. N. Lee, and R. Landick. 1995. Discontinuous movements of DNA and RNA in RNA polymerase accompany formation of a paused transcription complex. Cell 81:341–350.
267. Wang, Y., and P. H. von Hippel. 1993. Escherichia coli transcription termination factor rho. I. ATPase activation by oligonucleotide cofactors. J. Biol. Chem. 268:13940–13946.
268. Wang, Y., and P. H. von Hippel. 1993. Escherichia coli transcription termination factor rho. II. Binding of oligonucleotide cofactors. J. Biol. Chem. 268:13947–13955.
269. Ward, D. F., A. DeLong, and M. E. Gottesman. 1983. Escherichia coli nusB mutations that suppress nusA1 exhibit λ N specificity. J. Mol. Biol. 168:73–85.
270. Warren, F., and A. Das. 1984. Formation of termination-resistant transcription complex at phage lambda nut locus: effects of altered translation and a ribosomal mutation. Proc. Natl. Acad. Sci. USA 81:3612–3616.
271. Whalen, W., and A. Das. 1990. Action of an RNA site at a distance: role of the nut genetic signal in transcription antitermination by phage λN gene product. New Biol. 2:975–991.
272. Whalen, W., B. Ghosh, and A. Das. 1988. NusA protein is necessary and sufficient in vitro for phage λN gene product to suppress a ρ-independent terminator placed downstream of nutL. Proc. Natl. Acad. Sci. USA 85:2494–2498.
273. Wilson, K. S., and P. H. von Hipple. 1994. Stability of Escherichia coli transcription complexes near an intrinsic terminator. J. Mol. Biol. 244:36–51.
274. Wu, A. M., G. E. Christie, and T. Platt. 1981. Tandem termination sites in the tryptophan operon of Escherichia coli. Proc. Natl. Acad. Sci. USA 78:2913–2917.
275. Wu, S.-W., and T. Platt. 1993. Transcriptional arrest of yeast RNA polymerase II by Escherichia coli rho protein in vitro. Proc. Natl. Acad. Sci. USA 90:6606–6610.
276. Yager, T. D., and P. H. von Hipple. 1987. Transcript elongation and termination in Escherichia coli, p. 1241–1275. In F. C. Neidhardt, J. C. Ingraham, K. B. Low, B. Magasanik, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, vol. 2. American Society for Microbiology, Washington, D.C.
277. Yang, M. T., and J. F. Gardner. 1989. Transcription termination directed by heteroduplex thr attenuator templates. Evidence that the transcript stem and loop structure is the termination signal. J. Biol. Chem. 264:2634–2639.
278. Yang, X., J. A. Goliger, and J. W. Roberts. 1989. Specificity and mechanism of antitermination by Q proteins of bacteriophages λ and 82. J. Mol. Biol. 210:453–460.
279. Yang, X., C. M. Hart, E. J. Grayhack, and J. W. Roberts. 1987. Transcription antitermination by phage λ gene Q protein requires a DNA segment spanning the RNA start site. Genes Dev. 1:217–226.
280. Yang, X., and J. W. Roberts. 1989. Gene Q antiterminator proteins of Escherichia coli phages 82 and λ suppress pausing by RNA polymerase at a ρ-dependent terminator and other sites. Proc. Natl. Acad. Sci. USA 86:5301–5305.
281. Yanofsky, C., and V. Horn. 1981. Rifampin resistance mutations that alter the efficiency of transcription termination at the tryptophan operon attenuator. J. Bacteriol. 145:1334–1341.
282. Yarnell, W. S., and J. W. Roberts. 1992. The phage λ gene Q transcription antiterminator binds DNA in the late gene promoter as it modifies RNA polymerase. Cell 69:1181–1189.
283. Zalatan, F., J. Galloway-Salvo, and T. Platt. 1993. Deletion analysis of the Escherichia coli rho-dependent transcription terminator trp t'. J. Biol. Chem. 268:17051–17056.
284. Zalatan, F., and T. Platt. 1992. Effects of decreased cytosine content on Rho interaction with the Rho-dependent terminator trp t' in Escherichia coli. J. Biol. Chem. 267:19082–19088.
285. Zawel, L., and D. Reinberg. 1993. Initiation of transcription by RNA polymerase II: a multi-step process, p. 67–108. In W.E. Cohn and K. Moldave (ed.), Progress in Nucleic Acid Research and Molecular Biology. Academic Press, San Diego.
286. Zengel, J. M., and L. Lindahl. 1990. Escherichia coli ribosomal protein L4 stimulates transcription termination at a specific site in the leader of the S10 operon independent of L4-mediated inhibition of translation. J. Mol. Biol. 213:67–78.
287. Zengel, J. M., and L. Lindahl. 1990. Ribosomal protein L4 stimulates in vitro termination of transcription at a NusA-dependent terminator in the S10 operon leader. Proc. Natl. Acad. Sci. USA 87:2675–2679.
288. Zengel, J. M., and L. Lindahl. 1992. Ribosomal protein L4 and transcription factor NusA have separable roles in mediating termination of transcription within the leader of the S10 operon of Escherichia coli. Genes Dev. 6:2655–2662.
289. Zengel, J. M., and L. Lindahl. 1993. Domain I of 23S rRNA competes with a paused transcription complex for ribosomal protein L4 of Escherichia coli. Nucleic Acids Res. 21:2429–2435.
290. Zhang, Y., and M. M. Hanna. 1994. NusA changes the conformation of Escherichia coli RNA polymerase at the binding site for the 3' end of the nascent RNA. J. Bacteriol. 176:1787–1789.
291. Zheng, C., and D. I. Friedman. 1994. Reduced Rho-dependent transcription termination permits NusA-independent growth of Escherichia coli. Proc. Natl. Acad. Sci. USA 91:7543–7547.
292. Zuber, M., T. A. Patterson, and D. L. Court. 1987. Analysis of nutR, a site required for transcription antitermination in phage lambda. Proc. Natl. Acad. Sci. USA 84:4514–4518.
Chapter55
